{"noiseShare":0.4535,"topics":[{"claims":[187,276,291,335,419,472,531,459,594],"deltaPp":-2.7055,"id":0,"keywords":["segmentation","semantic","instance","medical","image"],"label":"Image Segmentation","labelFull":"Image Segmentation: Semantic, Instance, and Medical Imaging","papers":1755,"papersByYear":[96,136,153,163,201,252,249,218,287],"rel":-0.2759,"reviewed":true,"share":[9.8059,10.51,10.4366,9.8193,9.7055,10.7097,9.178,7.5932,7.1004],"size":3564,"slope":-0.003533},{"claims":[2,11,2,13,46,212,773,836,879],"deltaPp":12.9328,"id":1,"keywords":["diffusion","latent","text-to-image","denoising","process"],"label":"Efficient Diffusion Models","labelFull":"Efficient Diffusion Model Training and Inference","papers":1555,"papersByYear":[2,5,2,6,20,102,419,468,531],"rel":63.303,"reviewed":true,"share":[0.2043,0.3864,0.1364,0.3614,0.9657,4.3349,15.4442,16.3009,13.1371],"size":2774,"slope":0.022344},{"claims":[182,190,212,235,307,315,316,307,466],"deltaPp":-3.6676,"id":2,"keywords":["state-of-the-art","benchmark","outperforms","algorithm","optimization"],"label":"State-of-the-Art Benchmarking","labelFull":"State-of-the-Art Benchmark Performance Claims","papers":2105,"papersByYear":[134,149,165,183,261,272,274,262,405],"rel":-0.268,"reviewed":true,"share":[13.6874,11.5147,11.2551,11.0241,12.6026,11.5597,10.0995,9.1257,10.0198],"size":2530,"slope":-0.003935},{"claims":[224,189,254,264,232,311,249,220,177],"deltaPp":-6.245,"id":3,"keywords":["face","facial","recognition","expression","identity"],"label":"Face Representation Learning","labelFull":"Face Representation Learning for Recognition and Synthesis","papers":807,"papersByYear":[81,70,97,95,86,102,96,98,82],"rel":-0.7548,"reviewed":true,"share":[8.2737,5.4096,6.6166,5.7229,4.1526,4.3349,3.5385,3.4134,2.0287],"size":2120,"slope":-0.006419},{"claims":[145,150,178,185,206,174,160,203,194],"deltaPp":-3.3762,"id":4,"keywords":["depth","stereo","estimation","monocular","disparity"],"label":"Monocular Depth Estimation","labelFull":"Monocular Depth Estimation from Single Images","papers":730,"papersByYear":[58,65,82,71,92,91,80,88,103],"rel":-0.5699,"reviewed":true,"share":[5.9244,5.0232,5.5935,4.2771,4.4423,3.8674,2.9488,3.0651,2.5482],"size":1595,"slope":-0.00418},{"claims":[116,127,197,168,201,207,203,182,189],"deltaPp":-3.2836,"id":5,"keywords":["pose","estimation","human","camera","6d"],"label":"3D Pose Estimation","labelFull":"3D Pose Estimation for Humans, Objects, and Cameras","papers":843,"papersByYear":[60,68,85,85,110,109,118,93,115],"rel":-0.5358,"reviewed":true,"share":[6.1287,5.255,5.7981,5.1205,5.3114,4.6324,4.3494,3.2393,2.8451],"size":1590,"slope":-0.003761},{"claims":[88,130,198,132,211,218,192,182,199],"deltaPp":-2.4321,"id":6,"keywords":["adversarial","attacks","backdoor","perturbations","robustness"],"label":"Adversarial Robustness","labelFull":"Adversarial Attacks, Defenses, and Robustness in Computer Vision","papers":706,"papersByYear":[49,66,80,62,89,91,84,81,104],"rel":-0.4859,"reviewed":true,"share":[5.0051,5.1005,5.457,3.7349,4.2974,3.8674,3.0962,2.8213,2.573],"size":1550,"slope":-0.003526},{"claims":[93,108,151,194,205,193,166,120,154],"deltaPp":-2.2773,"id":7,"keywords":["detection","object","detectors","coco","weakly"],"label":"Object Detection","labelFull":"Object Detection Across Supervision Levels and Open-World","papers":784,"papersByYear":[47,65,76,93,124,112,92,73,102],"rel":-0.4744,"reviewed":true,"share":[4.8008,5.0232,5.1842,5.6024,5.9874,4.7599,3.3911,2.5427,2.5235],"size":1384,"slope":-0.003497},{"claims":[113,128,157,170,209,150,152,112,148],"deltaPp":-3.2924,"id":8,"keywords":["action","recognition","temporal","interaction","activity"],"label":"Video Action Recognition","labelFull":"Video Action Recognition and Temporal Action Detection","papers":668,"papersByYear":[55,59,68,79,101,76,73,63,94],"rel":-0.586,"reviewed":true,"share":[5.618,4.5595,4.6385,4.759,4.8769,3.2299,2.6907,2.1944,2.3256],"size":1339,"slope":-0.004282},{"claims":[57,97,102,162,219,204,147,154,148],"deltaPp":-1.7514,"id":9,"keywords":["label","class","loss","learning","samples"],"label":"Label-Efficient Learning","labelFull":"Label-Efficient Learning with Pseudo-Labels and Uncertainty","papers":884,"papersByYear":[45,63,72,106,142,136,106,99,115],"rel":-0.381,"reviewed":true,"share":[4.5965,4.8686,4.9113,6.3855,6.8566,5.7799,3.9071,3.4483,2.8451],"size":1290,"slope":-0.002313},{"claims":[57,70,81,127,151,105,94,131,170],"deltaPp":-0.8265,"id":10,"keywords":["video","temporal","frames","prediction","motion"],"label":"Scalable Video Modeling","labelFull":"Scalable Video Modeling and Efficient Temporal Learning","papers":721,"papersByYear":[42,45,55,81,100,88,71,99,140],"rel":-0.1927,"reviewed":true,"share":[4.2901,3.4776,3.7517,4.8795,4.8286,3.7399,2.617,3.4483,3.4636],"size":986,"slope":-0.001134},{"claims":[49,114,75,98,78,116,136,112,147],"deltaPp":-0.2959,"id":11,"keywords":["light","hyperspectral","illumination","reflection","imaging"],"label":"Intrinsic Image Decomposition","labelFull":"Decomposing Images into Reflectance and Illumination","papers":449,"papersByYear":[23,45,33,45,38,54,70,58,83],"rel":-0.126,"reviewed":true,"share":[2.3493,3.4776,2.251,2.7108,1.8349,2.2949,2.5802,2.0202,2.0534],"size":925,"slope":-0.000886},{"claims":[53,84,106,117,121,136,103,104,85],"deltaPp":-1.0166,"id":12,"keywords":["point","cloud","registration","3d","local"],"label":"Point Cloud Deep Learning","labelFull":"Deep Learning Architectures for 3D Point Cloud Analysis","papers":520,"papersByYear":[24,43,56,69,71,80,60,59,58],"rel":-0.4147,"reviewed":true,"share":[2.4515,3.323,3.8199,4.1566,3.4283,3.3999,2.2116,2.055,1.4349],"size":909,"slope":-0.001974},{"claims":[130,137,107,88,96,92,72,70,66],"deltaPp":-7.3001,"id":13,"keywords":["networks","convolutional","neural","layer","deep"],"label":"CNN Reparameterization","labelFull":"CNN Reparameterization and Drop-in Layer Replacement","papers":636,"papersByYear":[86,95,82,64,67,72,52,58,60],"rel":-0.831,"reviewed":true,"share":[8.7845,7.3416,5.5935,3.8554,3.2352,3.0599,1.9167,2.0202,1.4844],"size":858,"slope":-0.008886},{"claims":[91,97,57,54,57,78,89,112,183],"deltaPp":0.0034,"id":14,"keywords":["question","answering","vqa","visual","reasoning"],"label":"Visual Question Answering","labelFull":"Visual Question Answering (VQA) and Knowledge-Based Reasoning","papers":403,"papersByYear":[30,31,21,20,24,31,47,75,124],"rel":0.0011,"reviewed":true,"share":[3.0644,2.3957,1.4325,1.2048,1.1589,1.3175,1.7324,2.6123,3.0678],"size":818,"slope":0.000229},{"claims":[57,57,88,100,95,78,75,93,155],"deltaPp":-1.0852,"id":15,"keywords":["tracking","trackers","multi-object","siamese","visual"],"label":"Visual Object Tracking","labelFull":"Visual Object Tracking and Multi-Object Tracking Benchmarks","papers":372,"papersByYear":[30,20,39,40,37,40,42,44,80],"rel":-0.3541,"reviewed":true,"share":[3.0644,1.5456,2.6603,2.4096,1.7866,1.7,1.5481,1.5326,1.9792],"size":798,"slope":-0.001219},{"claims":[50,82,68,70,91,106,120,92,115],"deltaPp":-0.4012,"id":16,"keywords":["few-shot","zero-shot","learning","classification","generalized"],"label":"Zero/Few-Shot Learning","labelFull":"Zero-Shot and Few-Shot Learning for Image Classification","papers":491,"papersByYear":[25,40,41,40,53,67,76,62,87],"rel":-0.1571,"reviewed":true,"share":[2.5536,3.0912,2.7967,2.4096,2.5592,2.8474,2.8013,2.1595,2.1524],"size":794,"slope":-0.000659},{"claims":[6,34,42,90,79,127,128,84,164],"deltaPp":2.1891,"id":17,"keywords":["distillation","student","teacher","knowledge","loss"],"label":"Knowledge Distillation","labelFull":"Knowledge Distillation and Self-Distillation for Vision Models","papers":465,"papersByYear":[4,16,20,47,49,82,76,66,105],"rel":5.3576,"reviewed":true,"share":[0.4086,1.2365,1.3643,2.8313,2.366,3.4849,2.8013,2.2989,2.5977],"size":754,"slope":0.002579},{"claims":[39,59,64,63,79,44,120,90,176],"deltaPp":0.6665,"id":18,"keywords":["driving","autonomous","lane","vehicle","traffic"],"label":"End-to-End Autonomous Driving","labelFull":"End-to-End Autonomous Driving: Models, Datasets, Benchmarks","papers":359,"papersByYear":[16,24,33,28,33,27,59,46,93],"rel":0.4078,"reviewed":true,"share":[1.6343,1.8547,2.251,1.6867,1.5934,1.1475,2.1747,1.6022,2.3008],"size":734,"slope":0.000203},{"claims":[0,2,3,10,23,64,104,169,298],"deltaPp":5.2697,"id":19,"keywords":["vision-language","visual","benchmarks","learning","multiple"],"label":"Efficient Vision-Language Models","labelFull":"Efficiency and Compression for Vision-Language Models","papers":468,"papersByYear":[0,2,3,7,16,42,75,110,213],"rel":null,"reviewed":true,"share":[0.0,0.1546,0.2046,0.4217,0.7726,1.785,2.7645,3.8314,5.2697],"size":673,"slope":0.006432},{"claims":[42,50,47,24,64,107,105,99,127],"deltaPp":0.0328,"id":20,"keywords":["hand","grasp","pose","contact","3d"],"label":"3D Hand-Object Interaction","labelFull":"3D Hand-Object Interaction Modeling and Pose Estimation","papers":252,"papersByYear":[13,16,19,12,22,35,40,40,55],"rel":0.0247,"reviewed":true,"share":[1.3279,1.2365,1.296,0.7229,1.0623,1.4875,1.4744,1.3932,1.3607],"size":665,"slope":0.000287},{"claims":[6,10,15,47,72,88,131,153,135],"deltaPp":1.1564,"id":21,"keywords":["audio","sound","speech","music","visual"],"label":"Audio-Visual Learning","labelFull":"Audio-Visual Self-Supervised Learning and Generation","papers":235,"papersByYear":[2,6,6,19,27,31,39,50,55],"rel":5.6603,"reviewed":true,"share":[0.2043,0.4637,0.4093,1.1446,1.3037,1.3175,1.4375,1.7416,1.3607],"size":657,"slope":0.001781},{"claims":[5,16,29,35,55,59,85,106,223],"deltaPp":3.3767,"id":22,"keywords":["multimodal","modality","fusion","alignment","features"],"label":"Multimodal Fusion and Alignment","labelFull":"Multimodal Fusion and Alignment for Feature Integration","papers":436,"papersByYear":[4,13,19,25,39,41,63,79,153],"rel":8.2641,"reviewed":true,"share":[0.4086,1.0046,1.296,1.506,1.8831,1.7425,2.3222,2.7517,3.7853],"size":613,"slope":0.003506},{"claims":[46,73,98,87,79,79,64,37,37],"deltaPp":-2.3253,"id":23,"keywords":["imagenet","cifar-10","top-1","accuracy","cifar-100"],"label":"Image Classification","labelFull":"Image Classification on ImageNet and CIFAR Benchmarks","papers":484,"papersByYear":[31,53,77,71,66,65,55,32,34],"rel":-0.7343,"reviewed":true,"share":[3.1665,4.0958,5.2524,4.2771,3.1869,2.7624,2.0273,1.1146,0.8412],"size":600,"slope":-0.004368},{"claims":[6,12,14,19,19,18,26,82,397],"deltaPp":5.9249,"id":24,"keywords":["reasoning","vision-language","visual","spatial","chain-of-thought"],"label":"Vision-Language Reasoning","labelFull":"Vision-Language Reasoning with Multi-step Chain-of-Thought","papers":392,"papersByYear":[4,8,9,10,12,14,23,56,256],"rel":14.5005,"reviewed":true,"share":[0.4086,0.6182,0.6139,0.6024,0.5794,0.595,0.8478,1.9505,6.3335],"size":593,"slope":0.004693},{"claims":[33,43,54,55,86,76,53,86,100],"deltaPp":-0.8873,"id":25,"keywords":["attention","mechanism","visual","features","spatial"],"label":"Vision Attention Mechanisms","labelFull":"Efficient Deformable Attention for Vision Transformers","papers":482,"papersByYear":[30,37,41,36,68,65,46,71,88],"rel":-0.2896,"reviewed":true,"share":[3.0644,2.8594,2.7967,2.1687,3.2834,2.7624,1.6955,2.473,2.1771],"size":586,"slope":-0.001053},{"claims":[31,48,36,78,81,73,71,66,89],"deltaPp":-0.1587,"id":26,"keywords":["style","transfer","content","font","stylization"],"label":"Arbitrary Style Transfer","labelFull":"Arbitrary Image Style Transfer and Style Representation","papers":276,"papersByYear":[11,21,23,39,38,36,37,32,39],"rel":-0.1412,"reviewed":true,"share":[1.1236,1.6229,1.5689,2.3494,1.8349,1.53,1.3638,1.1146,0.9649],"size":573,"slope":-0.000565},{"claims":[29,27,43,43,60,67,74,63,164],"deltaPp":0.9443,"id":27,"keywords":["reconstruction","compressive","hdr","quality","sensing"],"label":"Neural Reconstruction","labelFull":"Neural Reconstruction for Imaging, Video, and 3D Vision","papers":429,"papersByYear":[22,19,30,34,45,47,51,52,129],"rel":0.4202,"reviewed":true,"share":[2.2472,1.4683,2.0464,2.0482,2.1729,1.9975,1.8798,1.8112,3.1915],"size":570,"slope":0.000737},{"claims":[25,69,56,87,70,92,56,71,40],"deltaPp":-0.5888,"id":28,"keywords":["domain","adaptation","source","generalization","target"],"label":"Domain Adaptation","labelFull":"Domain Adaptation Methods for Computer Vision","papers":352,"papersByYear":[14,34,31,45,45,59,45,45,34],"rel":-0.4117,"reviewed":true,"share":[1.43,2.6275,2.1146,2.7108,2.1729,2.5074,1.6587,1.5674,0.8412],"size":566,"slope":-0.001108},{"claims":[7,12,18,25,70,78,101,137,96],"deltaPp":1.2179,"id":29,"keywords":["rendering","inverse","relighting","differentiable","quality"],"label":"Efficient Neural Rendering","labelFull":"Efficient Neural Rendering: High-Quality Real-Time Synthesis","papers":390,"papersByYear":[6,7,10,19,49,61,76,88,74],"rel":1.9871,"reviewed":true,"share":[0.6129,0.541,0.6821,1.1446,2.366,2.5924,2.8013,3.0651,1.8308],"size":544,"slope":0.003022},{"claims":[72,57,70,63,66,36,51,41,71],"deltaPp":-2.3812,"id":30,"keywords":["person","re-identification","identity","reid","visible-infrared"],"label":"Person Re-Identification","labelFull":"Person Re-Identification (Image, Video, and Cross-Modality)","papers":256,"papersByYear":[33,28,32,33,27,16,25,22,40],"rel":-0.7064,"reviewed":true,"share":[3.3708,2.1638,2.1828,1.988,1.3037,0.68,0.9215,0.7663,0.9896],"size":527,"slope":-0.002925},{"claims":[37,72,57,63,57,59,49,65,61],"deltaPp":-1.3877,"id":31,"keywords":["flow","optical","scene","estimation","motion"],"label":"Optical Flow Estimation","labelFull":"Deep Learning for Optical Flow and Scene Flow","papers":275,"papersByYear":[24,37,32,26,26,28,27,32,43],"rel":-0.5661,"reviewed":true,"share":[2.4515,2.8594,2.1828,1.5663,1.2554,1.19,0.9952,1.1146,1.0638],"size":520,"slope":-0.002256},{"claims":[44,77,33,39,44,50,72,63,70],"deltaPp":-1.0286,"id":32,"keywords":["captioning","image","video","dense","generated"],"label":"Image and Video Captioning","labelFull":"Image and Video Captioning, Including Dense Generation","papers":237,"papersByYear":[20,24,15,18,20,25,38,36,41],"rel":-0.5035,"reviewed":true,"share":[2.0429,1.8547,1.0232,1.0843,0.9657,1.0625,1.4007,1.2539,1.0143],"size":492,"slope":-0.000864},{"claims":[26,26,44,44,66,60,53,76,90],"deltaPp":0.166,"id":33,"keywords":["compression","coding","rate-distortion","codecs","learned image"],"label":"Learned Image Compression","labelFull":"Neural Image Compression with Rate-Distortion Optimization","papers":230,"papersByYear":[10,10,17,19,29,33,30,34,48],"rel":0.1625,"reviewed":true,"share":[1.0215,0.7728,1.1596,1.1446,1.4003,1.4025,1.1058,1.1843,1.1875],"size":485,"slope":0.000341},{"claims":[17,8,11,13,13,28,40,70,283],"deltaPp":3.5492,"id":34,"keywords":["reinforcement","policy","reward","agent","gui"],"label":"RL Fine-Tuning LLMs","labelFull":"RL Fine-Tuning and Alignment of Large Language Models","papers":328,"papersByYear":[12,7,8,12,13,20,26,37,193],"rel":2.8957,"reviewed":true,"share":[1.2257,0.541,0.5457,0.7229,0.6277,0.85,0.9583,1.2887,4.7749],"size":483,"slope":0.002899},{"claims":[26,30,60,64,60,76,72,46,44],"deltaPp":-0.561,"id":35,"keywords":["super-resolution","low-resolution","image","video","network"],"label":"Super-Resolution","labelFull":"Image and Video Super-Resolution Methods","papers":248,"papersByYear":[13,14,27,33,26,36,40,28,31],"rel":-0.4225,"reviewed":true,"share":[1.3279,1.0819,1.8417,1.988,1.2554,1.53,1.4744,0.9753,0.7669],"size":478,"slope":-0.000626},{"claims":[1,1,2,0,3,5,49,132,262],"deltaPp":4.4254,"id":36,"keywords":["multimodal","reasoning","visual","capabilities","multi-modal"],"label":"Multimodal LLM Benchmarks","labelFull":"Benchmarking and Evaluating Multimodal Large Language Models","papers":316,"papersByYear":[1,1,2,0,2,5,35,87,183],"rel":43.3438,"reviewed":true,"share":[0.1021,0.0773,0.1364,0.0,0.0966,0.2125,1.2901,3.0303,4.5275],"size":455,"slope":0.004847},{"claims":[12,16,14,29,33,56,68,107,107],"deltaPp":0.5499,"id":37,"keywords":["anomaly","detection","industrial","normal","video"],"label":"Visual Anomaly Detection","labelFull":"Unsupervised Visual Anomaly Detection for Images and Videos","papers":184,"papersByYear":[6,9,8,11,12,22,27,42,47],"rel":0.8972,"reviewed":true,"share":[0.6129,0.6955,0.5457,0.6627,0.5794,0.935,0.9952,1.4629,1.1628],"size":442,"slope":0.000946},{"claims":[28,36,29,47,65,51,57,48,66],"deltaPp":-0.9017,"id":38,"keywords":["localization","keypoints","cross-view","visual","accuracy"],"label":"Visual Localization","labelFull":"Visual Localization and Cross-View Geo-Localization","papers":258,"papersByYear":[19,23,21,25,32,31,35,30,42],"rel":-0.4646,"reviewed":true,"share":[1.9408,1.7774,1.4325,1.506,1.5451,1.3175,1.2901,1.0449,1.0391],"size":427,"slope":-0.001046},{"claims":[5,15,30,30,70,72,55,68,81],"deltaPp":1.2554,"id":39,"keywords":["continual learning","incremental","forgetting","class","catastrophic"],"label":"Class-Incremental Learning","labelFull":"Class-Incremental Learning in Computer Vision","papers":246,"papersByYear":[2,9,17,21,33,35,31,39,59],"rel":6.1449,"reviewed":true,"share":[0.2043,0.6955,1.1596,1.2651,1.5934,1.4875,1.1426,1.3584,1.4597],"size":426,"slope":0.0012},{"claims":[21,38,26,36,39,44,52,71,98],"deltaPp":-0.0478,"id":40,"keywords":["retrieval","composed image","cross-modal","zero-shot","benchmarks"],"label":"Composed Image Retrieval","labelFull":"Composed Image Retrieval (Image+Text Queries)","papers":286,"papersByYear":[15,26,18,26,25,34,33,49,60],"rel":-0.0312,"reviewed":true,"share":[1.5322,2.0093,1.2278,1.5663,1.2071,1.445,1.2164,1.7067,1.4844],"size":425,"slope":-0.000207},{"claims":[24,38,37,54,48,31,68,41,46],"deltaPp":-0.434,"id":41,"keywords":["text","document","scene","recognition","layout"],"label":"Scene Text Spotting","labelFull":"End-to-End Scene Text Spotting and Recognition","papers":189,"papersByYear":[12,13,16,21,21,19,30,25,32],"rel":-0.3541,"reviewed":true,"share":[1.2257,1.0046,1.0914,1.2651,1.014,0.8075,1.1058,0.8708,0.7917],"size":387,"slope":-0.000428},{"claims":[16,16,37,46,39,65,61,57,47],"deltaPp":0.0798,"id":42,"keywords":["garment","clothing","human","mesh","body"],"label":"Clothed Human Reconstruction","labelFull":"3D Reconstruction and Modeling of Clothed Humans","papers":206,"papersByYear":[6,10,17,21,25,31,36,32,28],"rel":0.1302,"reviewed":true,"share":[0.6129,0.7728,1.1596,1.2651,1.2071,1.3175,1.3269,1.1146,0.6927],"size":384,"slope":0.000289},{"claims":[2,4,6,12,83,83,45,56,72],"deltaPp":1.1317,"id":43,"keywords":["vision","transformers","computer","foundation","architecture"],"label":"Vision Transformers","labelFull":"Vision Transformers: Efficient Architectures and Training","papers":266,"papersByYear":[2,3,5,12,49,61,36,44,54],"rel":5.5394,"reviewed":true,"share":[0.2043,0.2318,0.3411,0.7229,2.366,2.5924,1.3269,1.5326,1.336],"size":363,"slope":0.002045},{"claims":[4,8,9,16,32,61,87,81,65],"deltaPp":0.5595,"id":44,"keywords":["avatars","hair","head","animatable","high-fidelity"],"label":"3D Human Avatar Reconstruction","labelFull":"High-Fidelity Animatable 3D Human Avatar Reconstruction","papers":177,"papersByYear":[3,3,4,7,17,27,43,38,35],"rel":1.826,"reviewed":true,"share":[0.3064,0.2318,0.2729,0.4217,0.8209,1.1475,1.585,1.3236,0.8659],"size":363,"slope":0.001477},{"claims":[5,8,5,4,23,32,51,92,114],"deltaPp":1.7901,"id":45,"keywords":["generation","text-to-image","text-to-video","quality","autoregressive"],"label":"Text-to-Image Generation","labelFull":"Text-to-Image Generation, Editing, and Efficient Fine-Tuning","papers":267,"papersByYear":[5,6,5,4,14,23,42,75,93],"rel":3.5052,"reviewed":true,"share":[0.5107,0.4637,0.3411,0.241,0.676,0.9775,1.5481,2.6123,2.3008],"size":334,"slope":0.002793},{"claims":[24,40,27,36,44,46,36,39,40],"deltaPp":-0.5298,"id":46,"keywords":["quantization","quantized","post-training","activations","networks"],"label":"Post-Training Quantization","labelFull":"Post-Training Quantization of Deep Neural Networks","papers":149,"papersByYear":[11,15,13,16,16,18,18,18,24],"rel":-0.4715,"reviewed":true,"share":[1.1236,1.1592,0.8868,0.9639,0.7726,0.765,0.6635,0.627,0.5938],"size":332,"slope":-0.000727},{"claims":[8,19,35,41,47,61,44,36,33],"deltaPp":-0.0159,"id":47,"keywords":["3d object","object detection","monocular 3d","kitti","detectors"],"label":"3D Object Detection","labelFull":"3D Object Detection in Autonomous Driving and Indoor Scenes","papers":198,"papersByYear":[5,10,23,24,28,37,30,21,20],"rel":-0.0311,"reviewed":true,"share":[0.5107,0.7728,1.5689,1.4458,1.352,1.5725,1.1058,0.7315,0.4948],"size":324,"slope":-0.000164},{"claims":[6,12,17,22,39,24,60,65,78],"deltaPp":0.9769,"id":48,"keywords":["motion","human","capture","prediction","sequences"],"label":"3D Human Motion Prediction","labelFull":"3D Human Motion Prediction and Generation","papers":230,"papersByYear":[4,9,11,12,23,20,48,47,56],"rel":2.3908,"reviewed":true,"share":[0.4086,0.6955,0.7503,0.7229,1.1106,0.85,1.7693,1.6371,1.3855],"size":323,"slope":0.001483},{"claims":[0,0,0,1,21,36,44,63,157],"deltaPp":2.8699,"id":49,"keywords":["token","attention","information","introduces","mechanism"],"label":"Token Pruning","labelFull":"Token Pruning and Merging for Efficient Vision Transformers","papers":247,"papersByYear":[0,0,0,1,19,30,31,50,116],"rel":null,"reviewed":true,"share":[0.0,0.0,0.0,0.0602,0.9174,1.275,1.1426,1.7416,2.8699],"size":322,"slope":0.003367},{"claims":[13,16,13,17,17,37,44,62,102],"deltaPp":1.1341,"id":50,"keywords":["reconstruction","3d","feed-forward","multi-view","4d"],"label":"Image-Based 3D Reconstruction","labelFull":"Image-Based 3D Reconstruction from Single and Multi-View","papers":261,"papersByYear":[9,13,11,14,15,30,36,50,83],"rel":1.2337,"reviewed":true,"share":[0.9193,1.0046,0.7503,0.8434,0.7243,1.275,1.3269,1.7416,2.0534],"size":321,"slope":0.001389},{"claims":[3,6,4,21,34,46,59,76,72],"deltaPp":1.008,"id":51,"keywords":["view synthesis","novel-view","synthesis quality","scenes","rendering"],"label":"Neural Novel View Synthesis","labelFull":"Neural Rendering for High-Quality Novel View Synthesis","papers":242,"papersByYear":[2,4,4,12,27,41,41,62,49],"rel":4.9339,"reviewed":true,"share":[0.2043,0.3091,0.2729,0.7229,1.3037,1.7425,1.5112,2.1595,1.2123],"size":321,"slope":0.00218},{"claims":[16,19,18,28,26,38,50,51,71],"deltaPp":0.1565,"id":52,"keywords":["3d","geometry","geometric","2d","scene"],"label":"Self-Supervised 3D Representation Learning","labelFull":"Self-Supervised 3D Representation Learning From 2D Images","papers":268,"papersByYear":[13,17,16,21,21,32,46,42,60],"rel":0.1179,"reviewed":true,"share":[1.3279,1.3138,1.0914,1.2651,1.014,1.36,1.6955,1.4629,1.4844],"size":317,"slope":0.000396},{"claims":[2,14,14,29,42,62,53,45,53],"deltaPp":0.5132,"id":53,"keywords":["lidar","point clouds","3d object","detection","segmentation"],"label":"LiDAR 3D Perception","labelFull":"LiDAR-Based 3D Detection, Segmentation, and Mapping","papers":177,"papersByYear":[2,8,9,14,23,33,30,29,29],"rel":2.512,"reviewed":true,"share":[0.2043,0.6182,0.6139,0.8434,1.1106,1.4025,1.1058,1.0101,0.7175],"size":314,"slope":0.000795},{"claims":[41,18,58,34,44,41,24,18,35],"deltaPp":-1.5265,"id":54,"keywords":["shape","matching","correspondence","3d","implicit"],"label":"Non-Rigid Shape Matching","labelFull":"Non-Rigid 3D Shape Matching and Dense Correspondence","papers":199,"papersByYear":[21,15,33,20,29,25,16,15,25],"rel":-0.7117,"reviewed":true,"share":[2.145,1.1592,2.251,1.2048,1.4003,1.0625,0.5898,0.5225,0.6185],"size":313,"slope":-0.001913},{"claims":[25,30,34,42,48,38,32,25,38],"deltaPp":-1.2448,"id":55,"keywords":["video object","object segmentation","video segmentation","instance","panoptic"],"label":"Video Object Segmentation","labelFull":"Video Object Segmentation and Video Segmentation Tasks","papers":179,"papersByYear":[18,15,15,20,25,27,17,18,24],"rel":-0.677,"reviewed":true,"share":[1.8386,1.1592,1.0232,1.2048,1.2071,1.1475,0.6266,0.627,0.5938],"size":312,"slope":-0.001238},{"claims":[38,40,32,28,24,37,47,32,34],"deltaPp":-1.0063,"id":56,"keywords":["deblurring","blur","defocus","motion","kernel"],"label":"Image and Video Deblurring","labelFull":"Image and Video Deblurring Methods","papers":143,"papersByYear":[13,19,12,13,14,22,21,16,13],"rel":-0.7578,"reviewed":true,"share":[1.3279,1.4683,0.8186,0.7831,0.676,0.935,0.7741,0.5573,0.3216],"size":312,"slope":-0.001116},{"claims":[2,0,4,7,17,22,64,62,129],"deltaPp":1.676,"id":57,"keywords":["editing","image","complex","manipulation","enables"],"label":"Text-Guided Image Editing","labelFull":"Text-Guided Image Editing with Instructional Prompts and Benchmarks","papers":193,"papersByYear":[2,0,4,5,10,15,38,43,76],"rel":8.2036,"reviewed":true,"share":[0.2043,0.0,0.2729,0.3012,0.4829,0.6375,1.4007,1.4977,1.8803],"size":307,"slope":0.002298},{"claims":[0,0,1,6,52,55,74,58,57],"deltaPp":0.7669,"id":58,"keywords":["federated","client","learning","heterogeneity","aggregation"],"label":"Heterogeneous Federated Learning","labelFull":"Techniques For Heterogeneous Federated Learning","papers":134,"papersByYear":[0,0,1,3,21,24,30,24,31],"rel":null,"reviewed":true,"share":[0.0,0.0,0.0682,0.1807,1.014,1.02,1.1058,0.8359,0.7669],"size":303,"slope":0.001415},{"claims":[15,35,43,32,59,32,25,28,18],"deltaPp":-1.0589,"id":59,"keywords":["graph","matching","node","neural","edge"],"label":"Deep Graph Matching","labelFull":"Deep Graph Matching with Graph Neural Networks","papers":207,"papersByYear":[14,27,31,24,34,23,18,21,15],"rel":-0.7405,"reviewed":true,"share":[1.43,2.0866,2.1146,1.4458,1.6417,0.9775,0.6635,0.7315,0.3711],"size":287,"slope":-0.001945},{"claims":[2,3,8,10,16,26,28,72,118],"deltaPp":2.0256,"id":60,"keywords":["manipulation","robot","navigation","simulation","real-world"],"label":"Generalist Robotic Manipulation","labelFull":"Generalist Robotic Manipulation: Multi-Task and Sim2Real","papers":189,"papersByYear":[1,3,5,9,10,16,17,42,86],"rel":19.8394,"reviewed":true,"share":[0.1021,0.2318,0.3411,0.5422,0.4829,0.68,0.6266,1.4629,2.1277],"size":283,"slope":0.002084},{"claims":[5,23,27,17,33,31,38,70,38],"deltaPp":0.1573,"id":61,"keywords":["inpainting","outpainting","image","completion","video"],"label":"Neural Inpainting","labelFull":"Deep Learning-Based Image, Video, and 3D Inpainting","papers":158,"papersByYear":[5,11,16,10,16,16,23,34,27],"rel":0.308,"reviewed":true,"share":[0.5107,0.8501,1.0914,0.6024,0.7726,0.68,0.8478,1.1843,0.668],"size":282,"slope":0.000204},{"claims":[10,15,15,24,41,33,47,50,43],"deltaPp":0.3831,"id":62,"keywords":["scene graph","graph generation","genome","visual","3d"],"label":"Scene Graph Generation","labelFull":"Scene Graph Generation for Images, Videos, and 3D Scenes","papers":164,"papersByYear":[4,9,8,13,21,16,31,30,32],"rel":0.9376,"reviewed":true,"share":[0.4086,0.6955,0.5457,0.7831,1.014,0.68,1.1426,1.0449,0.7917],"size":278,"slope":0.000612},{"claims":[13,11,21,14,24,37,48,44,65],"deltaPp":0.621,"id":63,"keywords":["multi-task","task-specific","experts","transfer learning","parameters"],"label":"Parameter-Efficient Multi-Task Learning","labelFull":"Parameter-Efficient Multi-Task Learning in Computer Vision","papers":215,"papersByYear":[7,10,14,9,20,28,38,35,54],"rel":0.8685,"reviewed":true,"share":[0.715,0.7728,0.955,0.5422,0.9657,1.19,1.4007,1.2191,1.336],"size":277,"slope":0.000894},{"claims":[0,0,0,0,8,39,73,74,77],"deltaPp":1.5339,"id":64,"keywords":["prompts","tuning","text","textual","learning"],"label":"Prompt Tuning","labelFull":"Prompt Tuning for Vision-Language Models","papers":191,"papersByYear":[0,0,0,0,5,22,51,51,62],"rel":null,"reviewed":true,"share":[0.0,0.0,0.0,0.0,0.2414,0.935,1.8798,1.7764,1.5339],"size":271,"slope":0.002693},{"claims":[3,7,13,19,29,23,53,49,60],"deltaPp":0.6121,"id":65,"keywords":["mesh","vertices","deformation","geometry","surface"],"label":"Neural Mesh Generation","labelFull":"Neural Mesh Generation and Reconstruction","papers":163,"papersByYear":[2,6,11,11,21,18,31,30,33],"rel":2.9961,"reviewed":true,"share":[0.2043,0.4637,0.7503,0.6627,1.014,0.765,1.1426,1.0449,0.8164],"size":256,"slope":0.000846},{"claims":[0,5,6,1,12,28,73,67,56],"deltaPp":1.1381,"id":66,"keywords":["3d","generation","assets","editing","scenes"],"label":"Efficient 3D Generation","labelFull":"Fast High-Fidelity 3D Asset and Scene Generation","papers":181,"papersByYear":[0,3,6,1,11,20,49,45,46],"rel":null,"reviewed":true,"share":[0.0,0.2318,0.4093,0.0602,0.5311,0.85,1.8061,1.5674,1.1381],"size":248,"slope":0.002024},{"claims":[50,34,39,29,26,23,14,18,12],"deltaPp":-1.6407,"id":67,"keywords":["saliency","maps","salient","fixation","detection"],"label":"Saliency Detection","labelFull":"Visual Saliency Detection and Prediction in Images and Video","papers":120,"papersByYear":[18,19,14,15,11,12,8,15,8],"rel":-0.8924,"reviewed":true,"share":[1.8386,1.4683,0.955,0.9036,0.5311,0.51,0.2949,0.5225,0.1979],"size":245,"slope":-0.001852},{"claims":[10,7,13,9,26,26,42,48,63],"deltaPp":0.1477,"id":68,"keywords":["restoration","degradation","image","real-world","quality"],"label":"Universal Image Restoration","labelFull":"Universal Image Restoration on Real and Synthetic Degradations","papers":156,"papersByYear":[8,4,10,6,14,18,27,30,39],"rel":0.1807,"reviewed":true,"share":[0.8172,0.3091,0.6821,0.3614,0.676,0.765,0.9952,1.0449,0.9649],"size":244,"slope":0.000638},{"claims":[7,9,31,18,33,28,23,44,45],"deltaPp":0.4294,"id":69,"keywords":["trajectory","prediction","future","social","forecasting"],"label":"Trajectory Prediction","labelFull":"Trajectory Prediction for Multi-Agent Motion","papers":152,"papersByYear":[5,6,14,12,19,15,16,27,38],"rel":0.8408,"reviewed":true,"share":[0.5107,0.4637,0.955,0.7229,0.9174,0.6375,0.5898,0.9404,0.9401],"size":238,"slope":0.000389},{"claims":[6,7,12,22,38,38,24,25,63],"deltaPp":0.6552,"id":70,"keywords":["counterfactual","causal","explanations","intervention","confounders"],"label":"Counterfactual Explanations","labelFull":"Counterfactual Explanations for Vision Model Interpretability","papers":153,"papersByYear":[4,5,7,14,21,24,18,17,43],"rel":1.6035,"reviewed":true,"share":[0.4086,0.3864,0.4775,0.8434,1.014,1.02,0.6635,0.5921,1.0638],"size":235,"slope":0.000631},{"claims":[2,13,12,16,37,31,27,30,64],"deltaPp":0.6121,"id":71,"keywords":["navigation","vision-and-language","embodied","agents","environments"],"label":"Vision-and-Language Navigation","labelFull":"Vision-and-Language Navigation for Embodied Agents","papers":117,"papersByYear":[2,8,7,7,16,17,12,15,33],"rel":2.9961,"reviewed":true,"share":[0.2043,0.6182,0.4775,0.4217,0.7726,0.7225,0.4423,0.5225,0.8164],"size":232,"slope":0.000399},{"claims":[16,21,23,15,17,23,51,28,37],"deltaPp":-0.2944,"id":72,"keywords":["slam","odometry","visual","mapping","dense"],"label":"Dense Visual SLAM","labelFull":"Dense Visual SLAM and Neural Field Mapping","papers":102,"papersByYear":[7,13,12,6,7,9,19,12,17],"rel":-0.4117,"reviewed":true,"share":[0.715,1.0046,0.8186,0.3614,0.338,0.3825,0.7003,0.418,0.4206],"size":231,"slope":-0.000525},{"claims":[6,8,17,32,28,51,34,37,10],"deltaPp":-0.3375,"id":73,"keywords":["surface","signed","distance","implicit","neural"],"label":"Neural Implicit Surface Reconstruction","labelFull":"Neural Implicit Surface Reconstruction via SDF/UDF","papers":153,"papersByYear":[5,7,16,20,19,31,25,23,7],"rel":-0.6609,"reviewed":true,"share":[0.5107,0.541,1.0914,1.2048,0.9174,1.3175,0.9215,0.8011,0.1732],"size":223,"slope":-0.000133},{"claims":[33,42,19,20,26,31,13,20,17],"deltaPp":-1.1052,"id":74,"keywords":["crowd","counting","pedestrian","density","detection"],"label":"Crowd Counting & Detection","labelFull":"Crowd Counting and Dense Pedestrian Detection","papers":92,"papersByYear":[13,15,8,7,14,13,6,7,9],"rel":-0.8323,"reviewed":true,"share":[1.3279,1.1592,0.5457,0.4217,0.676,0.5525,0.2212,0.2438,0.2227],"size":221,"slope":-0.001281},{"claims":[23,22,25,7,18,35,29,24,32],"deltaPp":-0.2418,"id":75,"keywords":["sketch","image retrieval","fine-grained","strokes","photo"],"label":"Sketch-Based Image Retrieval","labelFull":"Sketch-Based Image Retrieval and Representation Learning","papers":83,"papersByYear":[6,8,7,4,7,14,13,9,15],"rel":-0.3945,"reviewed":true,"share":[0.6129,0.6182,0.4775,0.241,0.338,0.595,0.4792,0.3135,0.3711],"size":215,"slope":-0.000254},{"claims":[8,5,5,15,29,45,53,19,30],"deltaPp":0.0056,"id":76,"keywords":["query","entity","self-attention","cross-attention","features"],"label":"DETR Query Mechanisms","labelFull":"Query Mechanisms and Refinement in Detection Transformers (DETR)","papers":162,"papersByYear":[6,4,5,11,27,34,33,17,25],"rel":0.0091,"reviewed":true,"share":[0.6129,0.3091,0.3411,0.6627,1.3037,1.445,1.2164,0.5921,0.6185],"size":209,"slope":0.000567},{"claims":[1,4,8,12,17,11,34,41,81],"deltaPp":0.8133,"id":77,"keywords":["forgery","watermark","steganography","fake","localization"],"label":"Image Authentication","labelFull":"Image Forgery Detection, Localization, and Watermarking","papers":97,"papersByYear":[1,1,4,5,9,4,18,18,37],"rel":7.9657,"reviewed":true,"share":[0.1021,0.0773,0.2729,0.3012,0.4346,0.17,0.6635,0.627,0.9154],"size":209,"slope":0.000925},{"claims":[14,14,24,29,27,26,24,24,25],"deltaPp":-0.2697,"id":78,"keywords":["denoising","noise","noisy","image","self-supervised"],"label":"Self-Supervised Image Denoising","labelFull":"Self-Supervised Image Denoising Techniques","papers":116,"papersByYear":[7,8,12,14,15,12,16,14,18],"rel":-0.3772,"reviewed":true,"share":[0.715,0.6182,0.8186,0.8434,0.7243,0.51,0.5898,0.4876,0.4453],"size":207,"slope":-0.000377},{"claims":[9,18,14,13,34,18,24,33,39],"deltaPp":0.0152,"id":79,"keywords":["gaze","eye","estimation","target","3d"],"label":"Gaze Estimation","labelFull":"Eye Gaze Estimation, Target Detection, and Domain Adaptation","papers":64,"papersByYear":[3,7,4,4,8,6,8,11,13],"rel":0.0496,"reviewed":true,"share":[0.3064,0.541,0.2729,0.241,0.3863,0.255,0.2949,0.3831,0.3216],"size":202,"slope":-5.9e-05},{"claims":[1,8,8,7,18,17,33,40,66],"deltaPp":1.3823,"id":80,"keywords":["generative","quality","diversity","prior","sampling"],"label":"Guided Generation","labelFull":"Guided Generation for Quality, Diversity, and Efficiency","papers":167,"papersByYear":[1,7,7,6,15,13,25,33,60],"rel":13.5387,"reviewed":true,"share":[0.1021,0.541,0.4775,0.3614,0.7243,0.5525,0.9215,1.1494,1.4844],"size":198,"slope":0.001406},{"claims":[12,17,26,13,34,18,22,28,26],"deltaPp":-0.5825,"id":81,"keywords":["scenes","indoor","outdoor","occlusion","rgb-d"],"label":"Robust 3D Scene Understanding","labelFull":"Robust 3D Scene Understanding for Indoor and Outdoor","papers":181,"papersByYear":[12,15,20,13,32,18,20,25,26],"rel":-0.4752,"reviewed":true,"share":[1.2257,1.1592,1.3643,0.7831,1.5451,0.765,0.7372,0.8708,0.6432],"size":196,"slope":-0.000745},{"claims":[4,9,15,14,20,17,23,33,59],"deltaPp":1.0016,"id":82,"keywords":["inference","speed","faster","incurring","additional"],"label":"Efficient Inference","labelFull":"Efficient Inference for Real-Time Vision","papers":188,"papersByYear":[4,8,15,14,20,17,22,31,57],"rel":2.4513,"reviewed":true,"share":[0.4086,0.6182,1.0232,0.8434,0.9657,0.7225,0.8109,1.0798,1.4102],"size":194,"slope":0.000808},{"claims":[2,0,2,5,8,23,20,59,72],"deltaPp":1.0575,"id":83,"keywords":["video understanding","video-language","understanding benchmarks","long","multi-modal"],"label":"Long-Form Video Understanding","labelFull":"Multimodal LLMs for Long-Form Video Understanding","papers":136,"papersByYear":[2,0,2,5,5,17,15,39,51],"rel":5.1762,"reviewed":true,"share":[0.2043,0.0,0.1364,0.3012,0.2414,0.7225,0.5529,1.3584,1.2618],"size":191,"slope":0.001593},{"claims":[0,0,0,0,0,0,22,76,89],"deltaPp":1.7071,"id":84,"keywords":["gaussian","splatting","3d","primitives","scene"],"label":"3D Gaussian Splatting","labelFull":"Neural 3D Gaussian Splatting for Efficient Scene Representation","papers":144,"papersByYear":[0,0,0,0,0,0,17,58,69],"rel":null,"reviewed":true,"share":[0.0,0.0,0.0,0.0,0.0,0.0,0.6266,2.0202,1.7071],"size":187,"slope":0.002357},{"claims":[24,25,25,46,32,15,4,6,9],"deltaPp":-1.2568,"id":85,"keywords":["translation","image-to-image","unpaired","domains","latent"],"label":"Unpaired Image-to-Image Translation","labelFull":"Unpaired Image-to-Image Translation Using GANs","papers":115,"papersByYear":[14,17,17,22,19,9,4,6,7],"rel":-0.8789,"reviewed":true,"share":[1.43,1.3138,1.1596,1.3253,0.9174,0.3825,0.1474,0.209,0.1732],"size":186,"slope":-0.001885},{"claims":[7,9,5,16,15,18,33,33,48],"deltaPp":0.5284,"id":86,"keywords":["views","features","contrastive","attention","module"],"label":"Multi-View Contrastive Learning","labelFull":"Contrastive Multi-View Representation Learning","papers":144,"papersByYear":[5,8,5,12,14,13,22,23,42],"rel":1.0347,"reviewed":true,"share":[0.5107,0.6182,0.3411,0.7229,0.676,0.5525,0.8109,0.8011,1.0391],"size":184,"slope":0.000572},{"claims":[0,0,0,13,28,68,51,11,9],"deltaPp":0.1484,"id":87,"keywords":["radiance","fields","neural","rendering","nerf"],"label":"Neural Radiance Fields","labelFull":"Neural Radiance Fields: 3D Scene Modeling and Rendering","papers":135,"papersByYear":[0,0,0,9,20,51,39,10,6],"rel":null,"reviewed":true,"share":[0.0,0.0,0.0,0.5422,0.9657,2.1674,1.4375,0.3483,0.1484],"size":180,"slope":0.001023},{"claims":[19,35,31,31,22,29,5,4,3],"deltaPp":-1.0494,"id":88,"keywords":["gan","discriminator","generator","stylegan2","conditional"],"label":"GAN Stabilization","labelFull":"Stabilization and Generalization Techniques for GAN Training","papers":107,"papersByYear":[11,18,18,25,16,11,3,2,3],"rel":-0.934,"reviewed":true,"share":[1.1236,1.391,1.2278,1.506,0.7726,0.4675,0.1106,0.0697,0.0742],"size":179,"slope":-0.001906},{"claims":[8,2,5,7,21,33,28,35,39],"deltaPp":0.1229,"id":89,"keywords":["image-text","textual","contrastive","pairs","embeddings"],"label":"Contrastive Vision-Language Pretraining","labelFull":"Contrastive Vision-Language Pretraining for Image-Text Tasks","papers":160,"papersByYear":[8,2,5,5,18,26,26,32,38],"rel":0.1504,"reviewed":true,"share":[0.8172,0.1546,0.3411,0.3012,0.8691,1.105,0.9583,1.1146,0.9401],"size":178,"slope":0.000902},{"claims":[5,7,3,22,13,31,31,36,27],"deltaPp":0.3647,"id":90,"keywords":["layout","design","generation","elements","diverse"],"label":"Visual Layout Generation","labelFull":"Visual Layout Generation for Image Synthesis and UI Design","papers":105,"papersByYear":[2,3,3,13,10,12,15,24,23],"rel":1.7851,"reviewed":true,"share":[0.2043,0.2318,0.2046,0.7831,0.4829,0.51,0.5529,0.8359,0.569],"size":175,"slope":0.000616},{"claims":[0,1,1,0,4,1,23,48,92],"deltaPp":1.0638,"id":91,"keywords":["hallucination","multimodal","reduces","vision-language","mitigates"],"label":"Vision-Language Hallucination Mitigation","labelFull":"Hallucination Mitigation in Vision-Language Models","papers":81,"papersByYear":[0,1,1,0,2,1,12,21,43],"rel":null,"reviewed":true,"share":[0.0,0.0773,0.0682,0.0,0.0966,0.0425,0.4423,0.7315,1.0638],"size":170,"slope":0.001168},{"claims":[2,2,15,20,29,24,39,15,15],"deltaPp":-0.0064,"id":92,"keywords":["few-shot segmentation","few-shot semantic","semantic segmentation","support","instance"],"label":"Few-Shot Segmentation","labelFull":"Few-Shot Semantic and Instance Segmentation","papers":81,"papersByYear":[2,1,7,10,11,14,19,9,8],"rel":-0.0313,"reviewed":true,"share":[0.2043,0.0773,0.4775,0.6024,0.5311,0.595,0.7003,0.3135,0.1979],"size":161,"slope":0.000187},{"claims":[7,17,25,17,28,21,20,7,15],"deltaPp":-0.0837,"id":93,"keywords":["registration","diffeomorphic","recall","medical image","deformable"],"label":"Learning-Based 3D Registration","labelFull":"Deep Learning for 3D Image and Point Cloud Registration","papers":86,"papersByYear":[3,10,12,13,10,11,12,6,9],"rel":-0.2732,"reviewed":true,"share":[0.3064,0.7728,0.8186,0.7831,0.4829,0.4675,0.4423,0.209,0.2227],"size":157,"slope":-0.000516},{"claims":[2,7,11,22,9,28,23,20,30],"deltaPp":0.1948,"id":94,"keywords":["emotion","affective","recognition","multimodal","content"],"label":"Multimodal Emotion Recognition","labelFull":"Multimodal Emotion Recognition for Visual and Audio Content","papers":59,"papersByYear":[1,3,3,8,3,12,9,8,12],"rel":1.9079,"reviewed":true,"share":[0.1021,0.2318,0.2046,0.4819,0.1449,0.51,0.3317,0.2786,0.2969],"size":152,"slope":0.0002},{"claims":[18,23,11,28,9,5,10,25,22],"deltaPp":-0.2386,"id":95,"keywords":["rain","streaks","weather","removal","rainy"],"label":"Image Deraining","labelFull":"Image Deraining: Rain Streak and Raindrop Removal","papers":66,"papersByYear":[5,8,6,11,4,3,6,12,11],"rel":-0.4672,"reviewed":true,"share":[0.5107,0.6182,0.4093,0.6627,0.1931,0.1275,0.2212,0.418,0.2721],"size":151,"slope":-0.000411},{"claims":[1,1,3,6,22,22,51,18,26],"deltaPp":0.2937,"id":96,"keywords":["gesture","dance","co-speech","music","audio"],"label":"Audio-driven Gesture Generation","labelFull":"Audio-driven Gesture Generation for Co-speech and Music","papers":59,"papersByYear":[1,1,1,2,6,8,16,8,16],"rel":2.8766,"reviewed":true,"share":[0.1021,0.0773,0.0682,0.1205,0.2897,0.34,0.5898,0.2786,0.3958],"size":150,"slope":0.000507},{"claims":[9,7,8,13,18,26,18,18,31],"deltaPp":0.0056,"id":97,"keywords":["uncertainty","distribution","ensemble","calibration","epistemic"],"label":"Uncertainty Quantification","labelFull":"Predictive Uncertainty Quantification for Vision Models","papers":114,"papersByYear":[6,6,7,10,17,15,12,16,25],"rel":0.0091,"reviewed":true,"share":[0.6129,0.4637,0.4775,0.6024,0.8209,0.6375,0.4423,0.5573,0.6185],"size":148,"slope":4.5e-05},{"claims":[0,0,0,0,6,19,16,26,80],"deltaPp":1.4102,"id":98,"keywords":["tokens","visual","vision","pruning","transformers"],"label":"Vision Token Pruning","labelFull":"Token Pruning and Merging for Efficient Vision Transformers","papers":106,"papersByYear":[0,0,0,0,4,13,11,21,57],"rel":null,"reviewed":true,"share":[0.0,0.0,0.0,0.0,0.1931,0.5525,0.4055,0.7315,1.4102],"size":147,"slope":0.001533}],"totals":{"claims":[4868,6358,7291,8325,10354,11610,13530,14413,21044],"papers":[979,1294,1466,1660,2071,2353,2713,2871,4042]},"years":[2018,2019,2020,2021,2022,2023,2024,2025,2026]}