{"noiseShare":0.4709,"topics":[{"claims":[1,13,6,19,63,492,1409,1557,881],"deltaPp":13.0058,"id":0,"keywords":["diffusion","latent","text-to-image","generation","process"],"label":"Efficient Diffusion Models","labelFull":"Efficient Training and Sampling for Diffusion-Based Generation","papers":2453,"papersByYear":[1,9,5,8,29,241,747,885,528],"rel":228.1719,"reviewed":true,"share":[0.057,0.3799,0.1771,0.2445,0.7804,5.3449,14.6528,15.883,13.0628],"size":4441,"slope":0.022097},{"claims":[208,341,459,461,499,569,578,658,413],"deltaPp":-1.6189,"id":1,"keywords":["adversarial","attacks","robustness","perturbations","backdoor"],"label":"Adversarial Defense","labelFull":"Adversarial Defense and Training for Robust Vision Models","papers":1769,"papersByYear":[107,159,197,189,201,225,240,270,181],"rel":-0.2655,"reviewed":true,"share":[6.0969,6.7117,6.9759,5.7763,5.409,4.99,4.7077,4.8457,4.478],"size":4186,"slope":-0.002899},{"claims":[192,309,307,400,401,486,438,320,157],"deltaPp":-3.5711,"id":2,"keywords":["segmentation","semantic","instance","panoptic","supervised"],"label":"Semantic Segmentation","labelFull":"Semantic Segmentation (Weakly, Unsupervised, Panoptic)","papers":1811,"papersByYear":[110,178,190,235,238,298,250,203,109],"rel":-0.5698,"reviewed":true,"share":[6.2678,7.5137,6.728,7.1822,6.4047,6.609,4.9039,3.6432,2.6967],"size":3010,"slope":-0.00502},{"claims":[334,303,359,360,348,385,291,280,106],"deltaPp":-5.6703,"id":3,"keywords":["face","facial","expression","recognition","identity"],"label":"Face Representation Learning","labelFull":"Face Representation Learning for Recognition and Synthesis","papers":1134,"papersByYear":[126,119,143,134,137,151,135,128,61],"rel":-0.7898,"reviewed":true,"share":[7.1795,5.0232,5.0637,4.0954,3.6868,3.3489,2.6481,2.2972,1.5092],"size":2766,"slope":-0.006073},{"claims":[254,275,302,319,347,331,268,369,179],"deltaPp":-3.6154,"id":4,"keywords":["depth","stereo","estimation","monocular","disparity"],"label":"Monocular Depth Estimation","labelFull":"Monocular Depth Estimation from Single Images","papers":1208,"papersByYear":[106,112,143,129,153,164,137,166,98],"rel":-0.5986,"reviewed":true,"share":[6.0399,4.7277,5.0637,3.9425,4.1173,3.6372,2.6873,2.9792,2.4245],"size":2644,"slope":-0.004128},{"claims":[143,186,200,311,383,302,284,357,244],"deltaPp":-0.4351,"id":5,"keywords":["video","temporal","frames","event","features"],"label":"Efficient Video Transformers","labelFull":"Efficient Video Transformers Exploiting Temporal Redundancy","papers":1621,"papersByYear":[91,116,139,196,233,213,201,240,192],"rel":-0.0839,"reviewed":true,"share":[5.1852,4.8966,4.9221,5.9902,6.2702,4.7239,3.9427,4.3073,4.7501],"size":2410,"slope":-0.001122},{"claims":[156,232,287,331,391,352,290,212,141],"deltaPp":-2.2576,"id":6,"keywords":["object detection","detectors","coco","weakly","supervised"],"label":"Data-Efficient Object Detection","labelFull":"Data-Efficient Object Detection with Limited Supervision","papers":1326,"papersByYear":[80,127,149,165,203,205,165,139,93],"rel":-0.4953,"reviewed":true,"share":[4.5584,5.3609,5.2762,5.0428,5.4629,4.5465,3.2366,2.4946,2.3008],"size":2392,"slope":-0.003701},{"claims":[190,233,276,283,305,269,295,233,148],"deltaPp":-3.1047,"id":7,"keywords":["action","recognition","temporal","activity","egocentric"],"label":"Video Action Recognition","labelFull":"Video Action Recognition and Temporal Action Detection","papers":1108,"papersByYear":[94,113,119,134,148,145,142,122,91],"rel":-0.5797,"reviewed":true,"share":[5.3561,4.7699,4.2139,4.0954,3.9828,3.2158,2.7854,2.1895,2.2514],"size":2232,"slope":-0.003983},{"claims":[156,207,243,261,310,273,271,209,118],"deltaPp":-2.7351,"id":8,"keywords":["pose","estimation","human","6d","3d"],"label":"3D Pose Estimation","labelFull":"3D and 6D Pose Estimation for Humans and Objects","papers":1117,"papersByYear":[81,106,125,139,168,145,157,120,76],"rel":-0.5926,"reviewed":true,"share":[4.6154,4.4745,4.4263,4.2482,4.521,3.2158,3.0796,2.1536,1.8803],"size":2048,"slope":-0.003605},{"claims":[34,84,83,128,144,223,314,443,381],"deltaPp":3.0618,"id":9,"keywords":["medical","slide","image segmentation","pathology","classification"],"label":"Computational Pathology","labelFull":"Deep Learning for Computational Pathology Image Analysis","papers":757,"papersByYear":[14,37,42,51,66,96,128,167,156],"rel":3.8383,"reviewed":true,"share":[0.7977,1.5618,1.4873,1.5587,1.7761,2.1291,2.5108,2.9971,3.8595],"size":1834,"slope":0.003195},{"claims":[72,139,182,262,254,273,234,220,83],"deltaPp":-0.812,"id":10,"keywords":["point","cloud","registration","3d","completion"],"label":"Point Cloud Representation Learning","labelFull":"Deep Learning for 3D Point Cloud Representation and Analysis","papers":958,"papersByYear":[39,69,102,128,140,161,134,128,57],"rel":-0.3654,"reviewed":true,"share":[2.2222,2.9126,3.6119,3.912,3.7675,3.5706,2.6285,2.2972,1.4102],"size":1719,"slope":-0.001234},{"claims":[112,167,179,158,166,150,257,301,193],"deltaPp":-0.261,"id":11,"keywords":["driving","autonomous","pedestrian","lane","vehicle"],"label":"End-to-End Autonomous Driving","labelFull":"End-to-End Autonomous Driving: Models, Datasets, Simulation","papers":794,"papersByYear":[48,66,83,74,76,79,123,145,100],"rel":-0.0954,"reviewed":true,"share":[2.735,2.786,2.9391,2.2616,2.0452,1.7521,2.4127,2.6023,2.474],"size":1683,"slope":-0.000526},{"claims":[10,23,39,47,125,157,323,453,223],"deltaPp":2.9831,"id":12,"keywords":["rendering","gaussian","splatting","differentiable","inverse"],"label":"3D Gaussian Splatting","labelFull":"Efficient Real-Time 3D Gaussian Splatting for Neural Rendering","papers":887,"papersByYear":[8,14,28,35,84,111,194,274,139],"rel":6.5448,"reviewed":true,"share":[0.4558,0.591,0.9915,1.0697,2.2605,2.4617,3.8054,4.9174,3.4389],"size":1400,"slope":0.005322},{"claims":[137,113,138,170,173,172,146,174,152],"deltaPp":-1.2289,"id":13,"keywords":["tracking","trackers","multi-object","visual","benchmarks"],"label":"Visual Object Tracking","labelFull":"Visual Object Tracking and Multi-Object Tracking Benchmarks","papers":603,"papersByYear":[55,43,62,69,71,78,66,82,77],"rel":-0.3921,"reviewed":true,"share":[3.1339,1.8151,2.1955,2.1088,1.9107,1.7299,1.2946,1.4716,1.905],"size":1375,"slope":-0.001354},{"claims":[134,135,100,107,105,149,181,235,207],"deltaPp":0.7931,"id":14,"keywords":["question","answering","vqa","visual","reasoning"],"label":"Visual Question Answering","labelFull":"Visual Question Answering: Knowledge, Reasoning, Benchmarks","papers":667,"papersByYear":[46,45,38,42,42,64,99,153,138],"rel":0.3026,"reviewed":true,"share":[2.6211,1.8995,1.3456,1.2836,1.1302,1.4194,1.9419,2.7459,3.4142],"size":1353,"slope":0.001173},{"claims":[82,90,113,135,174,187,228,197,142],"deltaPp":-1.2507,"id":15,"keywords":["state-of-the-art","benchmark","outperforms","superior compared","experiments"],"label":"State-of-the-Art Benchmarking","labelFull":"Claims of State-of-the-Art Performance on Standard Benchmarks","papers":1301,"papersByYear":[81,90,109,129,169,183,218,186,136],"rel":-0.271,"reviewed":true,"share":[4.6154,3.7991,3.8598,3.9425,4.5479,4.0585,4.2762,3.3381,3.3647],"size":1348,"slope":-0.000906},{"claims":[65,128,132,136,175,165,194,170,112],"deltaPp":-0.0054,"id":16,"keywords":["few-shot","zero-shot","learning","classification","generalized"],"label":"Few-Shot and Zero-Shot Learning","labelFull":"Few-Shot and Zero-Shot Learning for Image Classification","papers":801,"papersByYear":[37,64,73,83,100,117,126,116,85],"rel":-0.0026,"reviewed":true,"share":[2.1083,2.7016,2.585,2.5367,2.6911,2.5948,2.4716,2.0818,2.1029],"size":1277,"slope":-0.000342},{"claims":[48,124,153,186,151,191,157,111,40],"deltaPp":-0.7468,"id":17,"keywords":["domain","adaptation","source","generalization","target"],"label":"Domain Adaptation","labelFull":"Domain Adaptation and Generalization for Vision Models","papers":703,"papersByYear":[27,61,84,101,90,126,109,73,32],"rel":-0.4854,"reviewed":true,"share":[1.5385,2.5749,2.9745,3.0868,2.422,2.7944,2.1381,1.3101,0.7917],"size":1161,"slope":-0.001458},{"claims":[36,39,64,90,137,181,200,251,148],"deltaPp":0.9964,"id":18,"keywords":["audio","sound","speech","music","visual"],"label":"Audio-Visual Representation Learning","labelFull":"Audio-Visual Representation Learning and Cross-Modal Alignment","papers":413,"papersByYear":[9,17,18,37,55,62,66,88,61],"rel":1.9431,"reviewed":true,"share":[0.5128,0.7176,0.6374,1.1308,1.4801,1.375,1.2946,1.5793,1.5092],"size":1146,"slope":0.001355},{"claims":[47,83,111,158,193,210,154,113,59],"deltaPp":-0.8885,"id":19,"keywords":["label","semi-supervised","pseudo","unlabeled","noisy"],"label":"Noisy-Label Semi-supervised Learning","labelFull":"Robust Semi-supervised Learning with Noisy Labels","papers":808,"papersByYear":[36,56,79,112,137,147,111,83,47],"rel":-0.4331,"reviewed":true,"share":[2.0513,2.3639,2.7975,3.423,3.6868,3.2601,2.1773,1.4896,1.1628],"size":1128,"slope":-0.001263},{"claims":[56,61,59,97,84,138,165,207,213],"deltaPp":1.8352,"id":20,"keywords":["reconstruction","quality","hdr","3d","images"],"label":"Neural 3D Reconstruction","labelFull":"Neural Implicit 3D Reconstruction from Images","papers":837,"papersByYear":[39,48,45,75,74,106,127,159,164],"rel":0.8258,"reviewed":true,"share":[2.2222,2.0262,1.5935,2.2922,1.9914,2.3509,2.4912,2.8536,4.0574],"size":1080,"slope":0.001946},{"claims":[62,112,110,140,94,149,156,149,99],"deltaPp":-0.4206,"id":21,"keywords":["illumination","light","imaging","reflection","polarization"],"label":"Intrinsic Image Decomposition","labelFull":"Intrinsic Image Decomposition and Illumination Estimation","papers":551,"papersByYear":[33,53,51,66,49,79,86,75,59],"rel":-0.2237,"reviewed":true,"share":[1.8803,2.2372,1.8059,2.0171,1.3186,1.7521,1.6869,1.346,1.4597],"size":1071,"slope":-0.00081},{"claims":[9,55,76,95,122,171,175,159,131],"deltaPp":1.8847,"id":22,"keywords":["distillation","knowledge","student","teacher","loss"],"label":"Knowledge Distillation","labelFull":"Knowledge Distillation Methods for Efficient Vision Models","papers":629,"papersByYear":[6,28,38,61,80,108,107,111,90],"rel":5.5124,"reviewed":true,"share":[0.3419,1.1819,1.3456,1.8643,2.1529,2.3952,2.0989,1.9921,2.2266],"size":993,"slope":0.002001},{"claims":[62,76,83,89,96,176,163,141,107],"deltaPp":-0.0263,"id":23,"keywords":["hand","grasp","pose","contact","interaction"],"label":"Hand-Object Pose Estimation","labelFull":"3D Hand-Object Pose Estimation and Interaction","papers":376,"papersByYear":[20,27,33,33,37,60,65,56,45],"rel":-0.0231,"reviewed":true,"share":[1.1396,1.1397,1.1686,1.0086,0.9957,1.3307,1.275,1.005,1.1133],"size":993,"slope":4e-06},{"claims":[52,88,72,142,134,121,125,158,96],"deltaPp":-0.1972,"id":24,"keywords":["style","transfer","stylization","content","font"],"label":"Arbitrary Style Transfer","labelFull":"Arbitrary Image Style Transfer and Stylization Methods","papers":440,"papersByYear":[23,34,33,61,58,56,59,71,45],"rel":-0.1505,"reviewed":true,"share":[1.3105,1.4352,1.1686,1.8643,1.5608,1.242,1.1573,1.2742,1.1133],"size":988,"slope":-0.000319},{"claims":[57,76,81,100,128,137,127,137,93],"deltaPp":-0.3967,"id":25,"keywords":["attention","spatial","mechanism","module","global"],"label":"Vision Attention Modules","labelFull":"Spatial, Channel, And Efficient Self-Attention for Vision","papers":748,"papersByYear":[43,56,61,72,98,108,111,116,83],"rel":-0.1619,"reviewed":true,"share":[2.4501,2.3639,2.1601,2.2005,2.6372,2.3952,2.1773,2.0818,2.0534],"size":936,"slope":-0.000367},{"claims":[81,139,87,81,83,100,149,135,76],"deltaPp":-0.9057,"id":26,"keywords":["captioning","image","video","dense","generated"],"label":"Image and Video Captioning","labelFull":"Image and Video Captioning: Diverse, Zero-Shot, Dense","papers":437,"papersByYear":[35,50,30,35,38,47,83,75,44],"rel":-0.4541,"reviewed":true,"share":[1.9943,2.1106,1.0623,1.0697,1.0226,1.0424,1.6281,1.346,1.0886],"size":931,"slope":-0.000802},{"claims":[79,101,107,110,133,114,84,131,69],"deltaPp":-1.5722,"id":27,"keywords":["flow","optical","scene","estimation","motion"],"label":"Optical Flow Estimation","labelFull":"Deep Learning for Optical Flow and Scene Flow Estimation","papers":517,"papersByYear":[48,60,62,52,65,54,54,75,47],"rel":-0.5748,"reviewed":true,"share":[2.735,2.5327,2.1955,1.5892,1.7492,1.1976,1.0592,1.346,1.1628],"size":928,"slope":-0.002086},{"claims":[10,31,42,62,82,100,157,230,213],"deltaPp":3.3295,"id":28,"keywords":["modality","multimodal","fusion","alignment","features"],"label":"Multimodal Fusion & Alignment","labelFull":"Multimodal Fusion and Alignment for Robust Representations","papers":652,"papersByYear":[8,24,30,47,62,66,110,152,153],"rel":7.3047,"reviewed":true,"share":[0.4558,1.0131,1.0623,1.4364,1.6685,1.4637,2.1577,2.7279,3.7853],"size":927,"slope":0.003447},{"claims":[121,124,136,141,98,65,71,87,69],"deltaPp":-1.9981,"id":29,"keywords":["person","re-identification","identity","visible-infrared","unsupervised"],"label":"Person Re-Identification","labelFull":"Person Re-Identification across Image, Video, and Infrared","papers":454,"papersByYear":[52,63,66,68,45,33,41,47,39],"rel":-0.6744,"reviewed":true,"share":[2.963,2.6593,2.3371,2.0782,1.211,0.7319,0.8042,0.8435,0.9649],"size":912,"slope":-0.002975},{"claims":[103,113,98,106,107,89,99,113,72],"deltaPp":-2.2193,"id":30,"keywords":["optimization","algorithm","convergence","gradient","matrix"],"label":"Efficient Optimization","labelFull":"Efficient Optimization Algorithms for Computer Vision","papers":622,"papersByYear":[65,71,71,67,66,62,79,81,60],"rel":-0.5992,"reviewed":true,"share":[3.7037,2.997,2.5142,2.0477,1.7761,1.375,1.5496,1.4537,1.4844],"size":900,"slope":-0.002685},{"claims":[36,61,111,120,128,133,103,88,43],"deltaPp":-0.4372,"id":31,"keywords":["super-resolution","low-resolution","image","video","network"],"label":"Super-Resolution","labelFull":"Deep Learning for Image and Video Super-Resolution","papers":456,"papersByYear":[22,32,51,64,65,70,60,59,33],"rel":-0.3488,"reviewed":true,"share":[1.2536,1.3508,1.8059,1.956,1.7492,1.5525,1.1769,1.0589,0.8164],"size":823,"slope":-0.000714},{"claims":[42,62,71,82,107,101,107,142,94],"deltaPp":0.1469,"id":32,"keywords":["compression","coding","jpeg","rate-distortion","codecs"],"label":"Learned Image Compression","labelFull":"Learned Image and Video Compression","papers":393,"papersByYear":[20,24,31,37,48,52,60,69,52],"rel":0.1289,"reviewed":true,"share":[1.1396,1.0131,1.0977,1.1308,1.2917,1.1532,1.1769,1.2383,1.2865],"size":808,"slope":0.000241},{"claims":[44,72,62,93,79,123,102,125,103],"deltaPp":-0.5916,"id":33,"keywords":["retrieval","composed image","cross-modal","image-text","benchmarks"],"label":"Cross-Modal Retrieval","labelFull":"Cross-Modal Retrieval for Image, Text, and Video","papers":517,"papersByYear":[36,43,41,55,52,86,63,82,59],"rel":-0.2884,"reviewed":true,"share":[2.0513,1.8151,1.4518,1.6809,1.3994,1.9073,1.2358,1.4716,1.4597],"size":803,"slope":-0.0006},{"claims":[130,138,116,70,101,80,64,56,29],"deltaPp":-4.3388,"id":34,"keywords":["convolutional","networks","neural","layer","deep"],"label":"Parameter-Efficient CNNs","labelFull":"Parameter-Efficient CNN Architectures","papers":604,"papersByYear":[87,100,91,60,79,68,50,44,25],"rel":-0.8752,"reviewed":true,"share":[4.9573,4.2212,3.2224,1.8337,2.1259,1.5081,0.9808,0.7897,0.6185],"size":784,"slope":-0.00541},{"claims":[11,21,37,56,75,103,146,195,105],"deltaPp":0.8532,"id":35,"keywords":["anomaly","detection","industrial","normal","video"],"label":"Visual Anomaly Detection","labelFull":"Unsupervised Image and Video Anomaly Detection","papers":299,"papersByYear":[5,11,18,20,26,41,59,73,46],"rel":2.9947,"reviewed":true,"share":[0.2849,0.4643,0.6374,0.6112,0.6997,0.9093,1.1573,1.3101,1.1381],"size":749,"slope":0.001215},{"claims":[58,83,113,89,112,85,71,86,49],"deltaPp":-1.1906,"id":36,"keywords":["graph","matching","node","feature","neural"],"label":"Deep Graph Matching","labelFull":"Deep Graph Matching Using Graph Neural Networks","papers":529,"papersByYear":[40,60,72,62,71,59,55,66,44],"rel":-0.5224,"reviewed":true,"share":[2.2792,2.5327,2.5496,1.8949,1.9107,1.3085,1.0789,1.1845,1.0886],"size":746,"slope":-0.002056},{"claims":[12,26,49,52,130,152,99,118,75],"deltaPp":0.9371,"id":37,"keywords":["continual learning","incremental","forgetting","catastrophic","class"],"label":"Continual Learning","labelFull":"Continual Learning for Vision: Reducing Forgetting","papers":398,"papersByYear":[7,13,34,28,63,81,57,61,54],"rel":2.3492,"reviewed":true,"share":[0.3989,0.5488,1.204,0.8557,1.6954,1.7964,1.1181,1.0948,1.336],"size":713,"slope":0.001026},{"claims":[0,6,8,14,26,85,122,238,211],"deltaPp":3.909,"id":38,"keywords":["vision-language","pre-trained","visual","benchmarks","learning"],"label":"Vision-Language Model Benchmarking","labelFull":"Benchmarking and Evaluation of Vision-Language Models","papers":516,"papersByYear":[0,3,7,11,21,68,88,160,158],"rel":null,"reviewed":true,"share":[0.0,0.1266,0.2479,0.3362,0.5651,1.5081,1.7262,2.8715,3.909],"size":710,"slope":0.004667},{"claims":[12,41,82,85,107,131,98,78,36],"deltaPp":0.0315,"id":39,"keywords":["3d object","object detection","monocular 3d","kitti","detectors"],"label":"Image-Based 3D Object Detection","labelFull":"Image-Based 3D Object Detection for Autonomous Driving","papers":418,"papersByYear":[9,23,52,51,67,87,60,47,22],"rel":0.0614,"reviewed":true,"share":[0.5128,0.9709,1.8414,1.5587,1.803,1.9295,1.1769,0.8435,0.5443],"size":670,"slope":-0.000202},{"claims":[56,76,80,82,111,75,80,64,41],"deltaPp":-0.5534,"id":40,"keywords":["text","document","scene","recognition","spotting"],"label":"Scene Text Recognition","labelFull":"Scene Text Recognition and Spotting in Natural Images","papers":273,"papersByYear":[21,22,32,30,40,34,31,37,26],"rel":-0.4625,"reviewed":true,"share":[1.1966,0.9287,1.1331,0.9169,1.0764,0.754,0.6081,0.664,0.6432],"size":665,"slope":-0.000703},{"claims":[71,63,102,78,97,90,50,49,43],"deltaPp":-1.48,"id":41,"keywords":["shape","matching","shapenet","3d","completion"],"label":"3D Shape Representation Learning","labelFull":"Neural 3D Shape Representation for Matching and Completion","papers":410,"papersByYear":[39,46,61,47,62,61,30,34,30],"rel":-0.666,"reviewed":true,"share":[2.2222,1.9417,2.1601,1.4364,1.6685,1.3528,0.5885,0.6102,0.7422],"size":643,"slope":-0.00219},{"claims":[50,69,71,68,108,117,56,61,42],"deltaPp":-1.2296,"id":42,"keywords":["video object","object segmentation","video segmentation","instance","referring"],"label":"Video Object Segmentation","labelFull":"Video Object Segmentation and Related Tasks","papers":364,"papersByYear":[32,36,37,36,56,69,34,40,24],"rel":-0.6743,"reviewed":true,"share":[1.8234,1.5196,1.3102,1.1002,1.507,1.5303,0.6669,0.7179,0.5938],"size":642,"slope":-0.001363},{"claims":[21,32,45,66,75,85,110,124,82],"deltaPp":0.6942,"id":43,"keywords":["motion","human","prediction","capture","sequences"],"label":"Human Motion Prediction","labelFull":"3D Human Motion Prediction and Generation from MoCap Data","papers":443,"papersByYear":[13,19,32,38,46,59,83,95,58],"rel":0.9372,"reviewed":true,"share":[0.7407,0.802,1.1331,1.1614,1.2379,1.3085,1.6281,1.705,1.4349],"size":640,"slope":0.001104},{"claims":[5,7,9,56,141,143,97,87,72],"deltaPp":1.1972,"id":44,"keywords":["vision","transformers","backbone","attention","architecture"],"label":"Efficient Vision Transformers","labelFull":"Efficient Vision Transformer Architectures and Training","papers":417,"papersByYear":[2,5,6,33,80,99,73,66,53],"rel":10.5018,"reviewed":true,"share":[0.114,0.2111,0.2125,1.0086,2.1529,2.1956,1.4319,1.1845,1.3112],"size":617,"slope":0.001889},{"claims":[19,62,48,57,69,74,97,147,41],"deltaPp":0.0584,"id":45,"keywords":["inpainting","outpainting","image","regions","video"],"label":"Image and Video Inpainting","labelFull":"Image, Video, and 3D Inpainting","papers":326,"papersByYear":[12,27,25,30,34,38,54,76,30],"rel":0.0854,"reviewed":true,"share":[0.6838,1.1397,0.8853,0.9169,0.915,0.8428,1.0592,1.364,0.7422],"size":614,"slope":0.000197},{"claims":[1,0,2,0,2,9,77,262,254],"deltaPp":4.2725,"id":46,"keywords":["multimodal","reasoning","multi-modal","capabilities","understanding"],"label":"Multimodal LLM Benchmarks","labelFull":"Benchmarks and Evaluation of Multimodal Large Language Models","papers":406,"papersByYear":[1,0,2,0,2,7,55,164,175],"rel":74.9561,"reviewed":true,"share":[0.057,0.0,0.0708,0.0,0.0538,0.1552,1.0789,2.9433,4.3295],"size":607,"slope":0.004682},{"claims":[0,0,0,1,17,102,194,188,104],"deltaPp":1.9792,"id":47,"keywords":["prompt","tuning","text","visual","textual"],"label":"Vision-Language Prompt Learning","labelFull":"Prompt Learning and Tuning for Vision-Language Models","papers":405,"papersByYear":[0,0,0,1,8,56,129,131,80],"rel":null,"reviewed":true,"share":[0.0,0.0,0.0,0.0306,0.2153,1.242,2.5304,2.351,1.9792],"size":606,"slope":0.00354},{"claims":[47,53,63,72,81,83,66,92,37],"deltaPp":-0.5383,"id":48,"keywords":["quantization","quantized","post-training","activations","error"],"label":"Post-Training Quantization","labelFull":"Post-Training Quantization Methods for Low-Bit Neural Networks","papers":265,"papersByYear":[19,21,29,29,35,33,32,45,22],"rel":-0.4972,"reviewed":true,"share":[1.0826,0.8864,1.0269,0.8863,0.9419,0.7319,0.6277,0.8076,0.5443],"size":594,"slope":-0.000557},{"claims":[44,77,78,100,93,81,50,41,25],"deltaPp":-1.3435,"id":49,"keywords":["imagenet","top-1","accuracy","resnet","higher"],"label":"Efficient ImageNet Classification","labelFull":"Efficient High-Accuracy ImageNet Classification","papers":497,"papersByYear":[34,61,64,84,79,69,43,39,24],"rel":-0.6935,"reviewed":true,"share":[1.9373,2.5749,2.2663,2.5672,2.1259,1.5303,0.8435,0.6999,0.5938],"size":589,"slope":-0.00248},{"claims":[35,70,69,104,96,85,53,51,25],"deltaPp":-0.9942,"id":50,"keywords":["loss","class","samples","negative","prototypes"],"label":"Class-Aware Losses","labelFull":"Class-Aware Losses for Discriminative Feature Learning","papers":467,"papersByYear":[27,54,48,82,75,70,44,45,22],"rel":-0.6462,"reviewed":true,"share":[1.5385,2.2794,1.6997,2.5061,2.0183,1.5525,0.8631,0.8076,0.5443],"size":588,"slope":-0.001836},{"claims":[31,19,26,28,27,56,64,96,238],"deltaPp":2.9696,"id":51,"keywords":["reinforcement","policy","reward","agent","learning"],"label":"RL Fine-Tuning","labelFull":"Reinforcement Learning for Model Fine-Tuning and Policy Optimization","papers":417,"papersByYear":[23,17,21,22,23,36,41,61,173],"rel":2.266,"reviewed":true,"share":[1.3105,0.7176,0.7436,0.6724,0.6189,0.7984,0.8042,1.0948,4.2801],"size":585,"slope":0.00221},{"claims":[6,29,43,61,73,90,63,116,98],"deltaPp":0.9693,"id":52,"keywords":["navigation","vision-and-language","embodied","agents","environments"],"label":"Vision-and-Language Navigation","labelFull":"Vision-and-Language Navigation for Embodied Agents","papers":302,"papersByYear":[6,18,22,33,31,46,35,58,53],"rel":2.835,"reviewed":true,"share":[0.3419,0.7598,0.779,1.0086,0.8342,1.0202,0.6865,1.0409,1.3112],"size":579,"slope":0.000758},{"claims":[101,108,100,82,55,41,30,30,20],"deltaPp":-1.6479,"id":53,"keywords":["saliency","salient","detection","object","maps"],"label":"Salient Object Detection","labelFull":"Salient Object Detection and Saliency Map Prediction","papers":247,"papersByYear":[35,42,38,36,25,20,17,20,14],"rel":-0.8263,"reviewed":true,"share":[1.9943,1.7729,1.3456,1.1002,0.6728,0.4436,0.3335,0.3589,0.3464],"size":567,"slope":-0.002252},{"claims":[0,0,0,11,46,81,87,174,163],"deltaPp":2.9441,"id":54,"keywords":["token","attention","introduces","information","mechanism"],"label":"Token Pruning","labelFull":"Token Pruning and Compression for Vision Transformers","papers":415,"papersByYear":[0,0,0,9,38,60,64,125,119],"rel":null,"reviewed":true,"share":[0.0,0.0,0.0,0.2751,1.0226,1.3307,1.2554,2.2434,2.9441],"size":562,"slope":0.003679},{"claims":[7,14,30,47,66,121,120,85,55],"deltaPp":0.482,"id":55,"keywords":["lidar","point","3d object","clouds","segmentation"],"label":"LiDAR 3D Perception","labelFull":"LiDAR-based 3D Detection and Semantic Segmentation","papers":312,"papersByYear":[5,9,20,28,37,66,65,51,31],"rel":1.6918,"reviewed":true,"share":[0.2849,0.3799,0.7082,0.8557,0.9957,1.4637,1.275,0.9153,0.7669],"size":545,"slope":0.000879},{"claims":[30,34,61,41,54,73,84,89,76],"deltaPp":0.0547,"id":56,"keywords":["memory","latency","computational","overhead","reduces"],"label":"Efficient Vision Models","labelFull":"Efficient Vision Models: Low Latency and Real-Time Inference","papers":496,"papersByYear":[29,33,52,36,51,66,76,84,69],"rel":0.0331,"reviewed":true,"share":[1.6524,1.393,1.8414,1.1002,1.3724,1.4637,1.4908,1.5075,1.7071],"size":542,"slope":3.7e-05},{"claims":[24,51,65,63,53,83,85,70,36],"deltaPp":-0.288,"id":57,"keywords":["clothing","garment","body","try-on","human"],"label":"Clothed Human Reconstruction","labelFull":"3D Reconstruction and Modeling of Clothed Humans","papers":266,"papersByYear":[12,27,23,32,32,40,46,38,16],"rel":-0.4212,"reviewed":true,"share":[0.6838,1.1397,0.8144,0.978,0.8611,0.8871,0.9023,0.682,0.3958],"size":530,"slope":-0.000407},{"claims":[33,51,54,86,77,72,56,61,15],"deltaPp":-1.1104,"id":58,"keywords":["visual","image classification","fine-grained","attribute","representation learning"],"label":"Self-supervised Fine-Grained Classification","labelFull":"Self-supervised Learning for Fine-Grained Classification","papers":428,"papersByYear":[26,46,48,65,64,64,48,52,15],"rel":-0.7495,"reviewed":true,"share":[1.4815,1.9417,1.6997,1.9866,1.7223,1.4194,0.9415,0.9332,0.3711],"size":505,"slope":-0.001592},{"claims":[49,55,59,55,70,61,78,43,35],"deltaPp":-0.6298,"id":59,"keywords":["deblurring","blur","defocus","motion","sharp"],"label":"Image Deblurring","labelFull":"Image Deblurring (Including Motion, Defocus, and Video)","papers":229,"papersByYear":[18,23,23,25,32,34,36,22,16],"rel":-0.6141,"reviewed":true,"share":[1.0256,0.9709,0.8144,0.7641,0.8611,0.754,0.7062,0.3948,0.3958],"size":505,"slope":-0.000746},{"claims":[3,3,10,18,31,37,123,132,144],"deltaPp":1.7836,"id":60,"keywords":["editing","image","instruction-based","video","consistency"],"label":"Instruction-based Image Editing","labelFull":"Instruction-based Image Editing with Benchmarks","papers":299,"papersByYear":[3,3,8,10,18,26,74,78,79],"rel":10.4365,"reviewed":true,"share":[0.1709,0.1266,0.2833,0.3056,0.4844,0.5766,1.4515,1.3999,1.9545],"size":501,"slope":0.00226},{"claims":[23,29,43,40,47,69,91,84,73],"deltaPp":0.6297,"id":61,"keywords":["multi-task","task-specific","transfer learning","gradient","parameters"],"label":"Parameter-Efficient Multi-Task Learning","labelFull":"Parameter-Efficient Multi-Task Learning for Vision Models","papers":374,"papersByYear":[15,21,29,31,37,53,64,64,60],"rel":0.7367,"reviewed":true,"share":[0.8547,0.8864,1.0269,0.9474,0.9957,1.1754,1.2554,1.1486,1.4844],"size":499,"slope":0.000665},{"claims":[57,97,72,98,84,50,21,5,3],"deltaPp":-1.8631,"id":62,"keywords":["gan","generator","discriminator","conditional","latent"],"label":"GAN Training Stabilization","labelFull":"Stabilizing GAN Training Across Architectures and Datasets","papers":292,"papersByYear":[34,57,43,58,54,27,13,3,3],"rel":-0.9617,"reviewed":true,"share":[1.9373,2.4061,1.5227,1.7726,1.4532,0.5988,0.255,0.0538,0.0742],"size":487,"slope":-0.003036},{"claims":[4,7,12,22,16,50,69,167,136],"deltaPp":1.9987,"id":63,"keywords":["affordance","manipulation","robotic","real-world","grounding"],"label":"Affordance-Based Robotic Manipulation","labelFull":"Affordance-Based Robotic Manipulation via Vision Models","papers":276,"papersByYear":[4,6,9,14,9,24,38,82,90],"rel":8.7701,"reviewed":true,"share":[0.2279,0.2533,0.3187,0.4279,0.2422,0.5323,0.7454,1.4716,2.2266],"size":483,"slope":0.002101},{"claims":[40,46,36,45,74,58,55,66,61],"deltaPp":-0.4521,"id":64,"keywords":["localization","cross-view","landmark","visual","location"],"label":"Cross-View Geo-Localization","labelFull":"Cross-View Visual Geo-Localization and Benchmarking","papers":303,"papersByYear":[24,34,25,26,41,36,39,41,37],"rel":-0.3306,"reviewed":true,"share":[1.3675,1.4352,0.8853,0.7946,1.1033,0.7984,0.765,0.7358,0.9154],"size":481,"slope":-0.000691},{"claims":[5,17,19,18,13,19,39,107,243],"deltaPp":3.9112,"id":65,"keywords":["reasoning","visual","spatial","chain-of-thought","understanding"],"label":"Visual Reasoning","labelFull":"Multi-step Visual Reasoning and Spatial Benchmarks","papers":333,"papersByYear":[3,12,11,10,9,16,28,79,165],"rel":22.8859,"reviewed":true,"share":[0.1709,0.5065,0.3895,0.3056,0.2422,0.3548,0.5492,1.4178,4.0821],"size":480,"slope":0.003125},{"claims":[3,3,14,18,36,65,130,128,61],"deltaPp":0.6455,"id":66,"keywords":["avatars","head","animatable","3d","high-fidelity"],"label":"3D Human Avatar Reconstruction","labelFull":"High-Fidelity Animatable 3D Human Avatar Reconstruction","papers":242,"papersByYear":[3,3,5,11,21,39,66,61,33],"rel":3.7771,"reviewed":true,"share":[0.1709,0.1266,0.1771,0.3362,0.5651,0.8649,1.2946,1.0948,0.8164],"size":458,"slope":0.001375},{"claims":[16,41,35,35,48,67,73,77,56],"deltaPp":0.4888,"id":67,"keywords":["3d","geometry","geometric","cad","2d"],"label":"Multi-View 3D Representation Learning","labelFull":"Self-Supervised Multi-View 3D Representation Pretraining","papers":377,"papersByYear":[14,27,29,28,40,57,64,66,52],"rel":0.6128,"reviewed":true,"share":[0.7977,1.1397,1.0269,0.8557,1.0764,1.2641,1.2554,1.1845,1.2865],"size":448,"slope":0.000493},{"claims":[16,33,41,79,60,50,69,69,29],"deltaPp":0.0959,"id":68,"keywords":["scene graph","graph generation","genome","visual grounding","relation"],"label":"Scene Graph Generation","labelFull":"Open-Vocabulary And Unbiased Scene Graph Generation","papers":240,"papersByYear":[7,18,19,40,32,25,39,40,20],"rel":0.2404,"reviewed":true,"share":[0.3989,0.7598,0.6728,1.2225,0.8611,0.5544,0.765,0.7179,0.4948],"size":446,"slope":-3.8e-05},{"claims":[9,14,11,33,41,75,93,106,61],"deltaPp":0.7789,"id":69,"keywords":["view synthesis","state-of-the-art view","synthesis quality","scenes","camera"],"label":"Novel View Synthesis","labelFull":"Neural Novel View Synthesis for Sparse and Single-View Inputs","papers":344,"papersByYear":[5,8,11,23,33,65,70,86,43],"rel":2.7339,"reviewed":true,"share":[0.2849,0.3377,0.3895,0.7029,0.8881,1.4416,1.3731,1.5434,1.0638],"size":443,"slope":0.001573},{"claims":[15,16,32,27,47,49,102,86,64],"deltaPp":0.2563,"id":70,"keywords":["restoration","degradation","image","real-world","quality"],"label":"All-in-One Image Restoration","labelFull":"All-in-One Image Restoration: Unified Multi-Task Restoration","papers":263,"papersByYear":[12,9,19,19,27,31,57,51,38],"rel":0.3748,"reviewed":true,"share":[0.6838,0.3799,0.6728,0.5807,0.7266,0.6875,1.1181,0.9153,0.9401],"size":438,"slope":0.000605},{"claims":[11,13,39,59,54,94,95,56,16],"deltaPp":-0.1515,"id":71,"keywords":["surface","signed","distance","implicit","reconstruction"],"label":"Neural Implicit Surface Reconstruction","labelFull":"Neural Implicit Surface Reconstruction From Multi-View Images","papers":285,"papersByYear":[7,11,29,38,35,59,60,36,10],"rel":-0.3798,"reviewed":true,"share":[0.3989,0.4643,1.0269,1.1614,0.9419,1.3085,1.1769,0.6461,0.2474],"size":437,"slope":6.4e-05},{"claims":[0,1,1,53,56,166,113,20,13],"deltaPp":0.2474,"id":72,"keywords":["radiance","fields","neural","rendering","nerf"],"label":"Neural Radiance Fields","labelFull":"Neural Radiance Field Modeling, Rendering, and Extensions","papers":309,"papersByYear":[0,1,1,32,42,121,86,16,10],"rel":null,"reviewed":true,"share":[0.0,0.0422,0.0354,0.978,1.1302,2.6835,1.6869,0.2872,0.2474],"size":423,"slope":0.001122},{"claims":[0,0,3,14,68,103,86,105,43],"deltaPp":0.668,"id":73,"keywords":["federated","learning","client","personalized","heterogeneity"],"label":"Heterogeneous Federated Learning","labelFull":"Federated Learning Under Data and Model Heterogeneity","papers":208,"papersByYear":[0,0,1,6,29,51,42,52,27],"rel":null,"reviewed":true,"share":[0.0,0.0,0.0354,0.1834,0.7804,1.1311,0.8239,0.9332,0.668],"size":422,"slope":0.001333},{"claims":[5,14,46,58,63,55,52,76,42],"deltaPp":0.7122,"id":74,"keywords":["trajectory","prediction","future","forecasting","diverse"],"label":"Trajectory Forecasting","labelFull":"Future Trajectory Forecasting for Agents in Autonomous Driving","papers":254,"papersByYear":[4,8,26,35,33,31,32,47,38],"rel":3.1251,"reviewed":true,"share":[0.2279,0.3377,0.9207,1.0697,0.8881,0.6875,0.6277,0.8435,0.9401],"size":411,"slope":0.000566},{"claims":[15,27,37,32,41,46,57,76,68],"deltaPp":0.8599,"id":75,"keywords":["inference","speed","faster","time","maintaining"],"label":"Real-Time Inference","labelFull":"Real-Time Inference and Speed Optimization for Vision Models","papers":383,"papersByYear":[14,24,37,30,41,44,55,71,67],"rel":1.078,"reviewed":true,"share":[0.7977,1.0131,1.3102,0.9169,1.1033,0.9758,1.0789,1.2742,1.6576],"size":399,"slope":0.000637},{"claims":[39,41,31,31,26,42,95,51,37],"deltaPp":-0.4664,"id":76,"keywords":["slam","odometry","visual","monocular","mapping"],"label":"Monocular Visual SLAM","labelFull":"Monocular Visual SLAM with Dense Mapping and Learning","papers":185,"papersByYear":[16,23,16,16,15,19,36,26,18],"rel":-0.5116,"reviewed":true,"share":[0.9117,0.9709,0.5666,0.489,0.4037,0.4214,0.7062,0.4666,0.4453],"size":393,"slope":-0.000528},{"claims":[0,0,0,7,25,42,72,116,130],"deltaPp":2.3008,"id":77,"keywords":["tokens","visual","vision","pruning","transformers"],"label":"Vision Token Pruning","labelFull":"Token Pruning and Compression for Vision Transformers","papers":282,"papersByYear":[0,0,0,5,20,32,49,83,93],"rel":null,"reviewed":true,"share":[0.0,0.0,0.0,0.1528,0.5382,0.7097,0.9612,1.4896,2.3008],"size":392,"slope":0.002692},{"claims":[6,15,20,32,46,75,83,62,50],"deltaPp":0.7467,"id":78,"keywords":["query","self-attention","support","decoder","cross-attention"],"label":"DETR Query Modeling","labelFull":"Query Modeling and Refinement for DETR Detection","papers":319,"papersByYear":[6,11,17,23,42,61,61,54,44],"rel":2.184,"reviewed":true,"share":[0.3419,0.4643,0.602,0.7029,1.1302,1.3528,1.1965,0.9691,1.0886],"size":389,"slope":0.001057},{"claims":[22,14,20,51,51,52,59,50,67],"deltaPp":0.4865,"id":79,"keywords":["explanations","counterfactual","causal","provides","generates"],"label":"Visual Counterfactual Explanations","labelFull":"Visual Counterfactual Explanations for Image Classification","papers":239,"papersByYear":[11,8,15,33,26,32,34,35,45],"rel":0.7762,"reviewed":true,"share":[0.6268,0.3377,0.5312,1.0086,0.6997,0.7097,0.6669,0.6281,1.1133],"size":386,"slope":0.000465},{"claims":[46,39,29,34,43,35,61,53,38],"deltaPp":-0.6148,"id":80,"keywords":["gaze","eye","estimation","target","human"],"label":"Eye Gaze Estimation","labelFull":"Deep Learning for Eye Gaze Estimation and Redirection","papers":134,"papersByYear":[16,15,10,11,15,13,22,20,12],"rel":-0.6743,"reviewed":true,"share":[0.9117,0.6332,0.3541,0.3362,0.4037,0.2883,0.4315,0.3589,0.2969],"size":378,"slope":-0.000529},{"claims":[23,24,47,53,40,52,53,52,26],"deltaPp":-0.2706,"id":81,"keywords":["denoising","noise","noisy","image","self-supervised"],"label":"Self-supervised Image Denoising","labelFull":"Self-supervised Image Denoising Techniques","papers":200,"papersByYear":[13,16,19,25,23,24,31,30,19],"rel":-0.3653,"reviewed":true,"share":[0.7407,0.6754,0.6728,0.7641,0.6189,0.5323,0.6081,0.5384,0.4701],"size":370,"slope":-0.000309},{"claims":[8,23,24,27,49,30,82,71,54],"deltaPp":0.3261,"id":82,"keywords":["mesh","vertex","vertices","topology","surface"],"label":"Neural Mesh Generation","labelFull":"Neural 3D Mesh Generation and Reconstruction","papers":235,"papersByYear":[6,16,18,18,34,25,49,42,27],"rel":0.9538,"reviewed":true,"share":[0.3419,0.6754,0.6374,0.5501,0.915,0.5544,0.9612,0.7538,0.668],"size":368,"slope":0.000365},{"claims":[49,29,44,18,48,44,47,42,32],"deltaPp":-0.3696,"id":83,"keywords":["sketch","image retrieval","fine-grained","photo","zero-shot"],"label":"Fine-Grained Sketch-Based Image Retrieval","labelFull":"Fine-Grained Sketch-Based Image Retrieval and Representation","papers":136,"papersByYear":[13,13,13,8,16,20,20,18,15],"rel":-0.499,"reviewed":true,"share":[0.7407,0.5488,0.4603,0.2445,0.4306,0.4436,0.3923,0.323,0.3711],"size":353,"slope":-0.000349},{"claims":[40,42,45,41,38,44,48,47,7],"deltaPp":-1.0482,"id":84,"keywords":["camera","calibration","epipolar","relative pose","minimal"],"label":"Minimal Camera Pose Solvers","labelFull":"Minimal Solvers for Camera Pose and Intrinsic Calibration","papers":205,"papersByYear":[21,22,26,24,18,30,29,29,6],"rel":-0.876,"reviewed":true,"share":[1.1966,0.9287,0.9207,0.7335,0.4844,0.6653,0.5689,0.5205,0.1484],"size":352,"slope":-0.001032},{"claims":[6,12,14,25,31,96,58,71,38],"deltaPp":0.2249,"id":85,"keywords":["lip","talking","head","speech","facial"],"label":"Audio-Driven Talking Heads","labelFull":"Audio-Driven Talking Head Generation and Control","papers":156,"papersByYear":[3,6,8,14,14,37,23,35,16],"rel":1.316,"reviewed":true,"share":[0.1709,0.2533,0.2833,0.4279,0.3767,0.8206,0.4512,0.6281,0.3958],"size":351,"slope":0.000459},{"claims":[26,35,54,43,50,61,44,20,12],"deltaPp":-1.0136,"id":86,"keywords":["cifar-10","cifar-100","imagenet","cifar100","cifar10"],"label":"Standard Image Classification Benchmarks","labelFull":"Image Classification on CIFAR and ImageNet Benchmarks","papers":328,"papersByYear":[23,34,52,39,50,55,43,20,12],"rel":-0.7734,"reviewed":true,"share":[1.3105,1.4352,1.8414,1.1919,1.3455,1.2198,0.8435,0.3589,0.2969],"size":345,"slope":-0.001542},{"claims":[50,43,57,70,57,26,19,12,9],"deltaPp":-1.3083,"id":87,"keywords":["translation","image-to-image","unpaired","unsupervised","domains"],"label":"Unpaired Image-to-Image Translation","labelFull":"Unpaired Image-to-Image Translation Approaches","papers":210,"papersByYear":[26,29,39,35,34,16,14,10,7],"rel":-0.8831,"reviewed":true,"share":[1.4815,1.2241,1.381,1.0697,0.915,0.3548,0.2746,0.1795,0.1732],"size":343,"slope":-0.001882},{"claims":[8,16,19,37,20,50,60,71,43],"deltaPp":0.5068,"id":88,"keywords":["layout","design","generation","graphic","elements"],"label":"Visual Layout Generation","labelFull":"Neural Visual Layout Generation for Graphic and UI Design","papers":178,"papersByYear":[5,8,14,16,13,20,28,42,32],"rel":1.7789,"reviewed":true,"share":[0.2849,0.3377,0.4958,0.489,0.3498,0.4436,0.5492,0.7538,0.7917],"size":324,"slope":0.000556},{"claims":[4,3,2,1,22,34,77,106,73],"deltaPp":1.2813,"id":89,"keywords":["text-to-image generation","text-to-video generation","prompts","evaluating","enables"],"label":"Text-to-Image Generation","labelFull":"Text-to-Image and Text-to-Video Generation","papers":262,"papersByYear":[4,1,2,1,12,23,64,94,61],"rel":5.6222,"reviewed":true,"share":[0.2279,0.0422,0.0708,0.0306,0.3229,0.5101,1.2554,1.687,1.5092],"size":322,"slope":0.002151},{"claims":[5,13,20,24,32,35,58,68,57],"deltaPp":1.0339,"id":90,"keywords":["generative","quality","diversity","prior","latent"],"label":"Generative Model Benchmarking","labelFull":"Comprehensive Benchmarking and Quality Assessment of Generative Models","papers":264,"papersByYear":[4,11,19,21,25,29,47,57,51],"rel":4.5366,"reviewed":true,"share":[0.2279,0.4643,0.6728,0.6418,0.6728,0.6432,0.9219,1.023,1.2618],"size":312,"slope":0.001052},{"claims":[20,31,32,50,44,34,34,28,23],"deltaPp":-0.5631,"id":91,"keywords":["patch","images","augmentation","synthetic","real"],"label":"Synthetic-to-Real Transfer","labelFull":"Synthetic-to-Real Transfer Learning for Vision Models","papers":255,"papersByYear":[19,28,27,41,36,30,28,25,21],"rel":-0.5201,"reviewed":true,"share":[1.0826,1.1819,0.9561,1.2531,0.9688,0.6653,0.5492,0.4487,0.5195],"size":296,"slope":-0.000976},{"claims":[18,30,33,33,36,42,29,46,26],"deltaPp":-0.3502,"id":92,"keywords":["occlusion","scenes","indoor","outdoor","occluded"],"label":"Occlusion-Aware Scene Understanding","labelFull":"Occlusion-Aware Scene Understanding in Indoor/Outdoor Scenes","papers":267,"papersByYear":[17,27,26,29,33,40,29,41,25],"rel":-0.3615,"reviewed":true,"share":[0.9687,1.1397,0.9207,0.8863,0.8881,0.8871,0.5689,0.7358,0.6185],"size":293,"slope":-0.000553},{"claims":[16,32,27,31,25,49,51,34,23],"deltaPp":-0.0773,"id":93,"keywords":["hyperspectral","compressive","spectral","sensing","snapshot"],"label":"Hyperspectral Image Reconstruction","labelFull":"Deep Learning for Compressive Hyperspectral Reconstruction","papers":139,"papersByYear":[7,14,12,16,16,20,25,16,13],"rel":-0.1938,"reviewed":true,"share":[0.3989,0.591,0.4249,0.489,0.4306,0.4436,0.4904,0.2872,0.3216],"size":288,"slope":-0.000189},{"claims":[6,23,18,20,36,43,68,40,33],"deltaPp":0.3261,"id":94,"keywords":["segmentation","3d instance","3d semantic","scannet","s3dis"],"label":"3D Segmentation","labelFull":"3D Semantic and Instance Segmentation","papers":212,"papersByYear":[6,18,15,16,25,31,43,31,27],"rel":0.9538,"reviewed":true,"share":[0.3419,0.7598,0.5312,0.489,0.6728,0.6875,0.8435,0.5564,0.668],"size":287,"slope":0.000253},{"claims":[29,34,27,40,21,30,28,51,24],"deltaPp":-0.1912,"id":95,"keywords":["rain","weather","rainy","streaks","removal"],"label":"Rain Removal","labelFull":"Rain Removal and Adverse Weather Image Restoration","papers":123,"papersByYear":[9,12,13,17,9,11,15,24,13],"rel":-0.3729,"reviewed":true,"share":[0.5128,0.5065,0.4603,0.5196,0.2422,0.244,0.2942,0.4307,0.3216],"size":284,"slope":-0.000267},{"claims":[11,11,13,29,39,52,44,36,40],"deltaPp":0.3359,"id":96,"keywords":["uncertainty","distribution","epistemic","calibration","shifts"],"label":"Uncertainty Quantification","labelFull":"Predictive Uncertainty Estimation and Calibration in Vision","papers":226,"papersByYear":[8,8,12,23,36,40,36,31,32],"rel":0.7369,"reviewed":true,"share":[0.4558,0.3377,0.4249,0.7029,0.9688,0.8871,0.7062,0.5564,0.7917],"size":275,"slope":0.000458},{"claims":[1,14,13,11,7,26,71,88,43],"deltaPp":0.4873,"id":97,"keywords":["occupancy","prediction","3d","forecasting","semantic"],"label":"3D Occupancy Prediction","labelFull":"3D Occupancy Prediction, Completion, and Forecasting","papers":138,"papersByYear":[1,4,7,9,5,14,35,41,22],"rel":8.5491,"reviewed":true,"share":[0.057,0.1688,0.2479,0.2751,0.1346,0.3105,0.6865,0.7358,0.5443],"size":274,"slope":0.00076},{"claims":[6,4,3,4,9,6,30,99,113],"deltaPp":1.7018,"id":98,"keywords":["video generation","quality","motion","benchmark","consistency"],"label":"Neural Video Generation","labelFull":"Neural Video Generation: Models, Benchmarks, and Quality","papers":197,"papersByYear":[4,3,3,4,6,6,19,74,78],"rel":7.4673,"reviewed":true,"share":[0.2279,0.1266,0.1062,0.1222,0.1615,0.1331,0.3727,1.3281,1.9297],"size":274,"slope":0.001826},{"claims":[1,1,15,23,36,48,53,49,47],"deltaPp":0.4378,"id":99,"keywords":["spiking","neural networks","neuromorphic","membrane","neuron"],"label":"Efficient Spiking Neural Networks","labelFull":"Efficient Training and Inference for Spiking Neural Networks","papers":108,"papersByYear":[1,1,5,10,13,19,19,20,20],"rel":7.6807,"reviewed":true,"share":[0.057,0.0422,0.1771,0.3056,0.3498,0.4214,0.3727,0.3589,0.4948],"size":273,"slope":0.000535},{"claims":[10,26,27,30,41,35,48,42,10],"deltaPp":-0.2579,"id":100,"keywords":["keypoint","2d","detector","3d","localization"],"label":"Keypoint Detection","labelFull":"2D/3D Keypoint Detection and Unsupervised Discovery","papers":169,"papersByYear":[8,16,21,18,24,25,29,20,8],"rel":-0.5658,"reviewed":true,"share":[0.4558,0.6754,0.7436,0.5501,0.6459,0.5544,0.5689,0.3589,0.1979],"size":269,"slope":-0.000388},{"claims":[9,15,19,28,20,42,47,45,43],"deltaPp":0.4596,"id":101,"keywords":["views","appearance","viewpoints","consistency","information"],"label":"Multi-View Representation Learning","labelFull":"View-Consistent Representations via Cross-View Fusion","papers":215,"papersByYear":[8,12,16,24,18,31,34,35,37],"rel":1.0083,"reviewed":true,"share":[0.4558,0.5065,0.5666,0.7335,0.4844,0.6875,0.6669,0.6281,0.9154],"size":268,"slope":0.000393},{"claims":[3,15,30,41,36,43,51,26,16],"deltaPp":0.2076,"id":102,"keywords":["student","teacher","knowledge","networks","pseudo labels"],"label":"Teacher-Student Distillation","labelFull":"Teacher-Student Knowledge Distillation via Pseudo-Labels","papers":190,"papersByYear":[2,9,24,26,26,32,36,22,13],"rel":1.8211,"reviewed":true,"share":[0.114,0.3799,0.8499,0.7946,0.6997,0.7097,0.7062,0.3948,0.3216],"size":261,"slope":8.4e-05},{"claims":[8,3,9,9,32,61,46,54,35],"deltaPp":0.4423,"id":103,"keywords":["image-text","pairs","alignment","contrastive","embeddings"],"label":"Contrastive Text-to-Image Alignment","labelFull":"Contrastive Alignment for Text-to-Image Generation","papers":220,"papersByYear":[7,3,8,7,27,44,38,52,34],"rel":1.1088,"reviewed":true,"share":[0.3989,0.1266,0.2833,0.2139,0.7266,0.9758,0.7454,0.9332,0.8412],"size":257,"slope":0.000979},{"claims":[2,10,27,38,54,36,47,28,14],"deltaPp":0.0592,"id":104,"keywords":["few-shot segmentation","few-shot semantic","semantic segmentation","pascal-5i","coco-20i"],"label":"Few-Shot Segmentation","labelFull":"Few-Shot Semantic Segmentation on PASCAL-5i & COCO-20i","papers":133,"papersByYear":[2,6,14,17,24,21,26,16,7],"rel":0.5193,"reviewed":true,"share":[0.114,0.2533,0.4958,0.5196,0.6459,0.4657,0.51,0.2872,0.1732],"size":256,"slope":5.2e-05},{"claims":[32,60,32,28,22,33,23,13,12],"deltaPp":-0.6493,"id":105,"keywords":["crowd","counting","density","localization","map"],"label":"Crowd Counting","labelFull":"Crowd Counting with Density Map Regression","papers":116,"papersByYear":[14,26,12,14,13,12,11,8,6],"rel":-0.814,"reviewed":true,"share":[0.7977,1.0975,0.4249,0.4279,0.3498,0.2661,0.2158,0.1436,0.1484],"size":255,"slope":-0.001006},{"claims":[4,31,78,56,32,14,17,12,10],"deltaPp":0.1087,"id":106,"keywords":["architecture search","neural architecture","search space","evolutionary","differentiable"],"label":"Neural Architecture Search","labelFull":"Neural Architecture Search: Efficient and Multi-Objective","papers":164,"papersByYear":[2,19,45,37,24,10,10,8,9],"rel":0.9535,"reviewed":true,"share":[0.114,0.802,1.5935,1.1308,0.6459,0.2218,0.1962,0.1436,0.2227],"size":254,"slope":-0.000874}],"totals":{"claims":[8586,11664,13919,16314,18486,22305,25271,28233,21044],"papers":[1755,2369,2824,3272,3716,4509,5098,5572,4042]},"years":[2018,2019,2020,2021,2022,2023,2024,2025,2026]}