{"noiseShare":0.4378,"topics":[{"claims":[90,null,198,null,262,null,337],"deltaPp":1.5355,"id":0,"keywords":["segmentation","semantic","instance","state-of-the-art","supervised"],"label":"Image Segmentation","labelFull":"Semantic, Instance, and Panoptic Image Segmentation","papers":465,"papersByYear":[46,null,108,null,133,null,178],"rel":0.259,"reviewed":true,"share":[5.9278,null,7.9529,null,8.0851,null,7.4633],"size":887,"slope":0.002369},{"claims":[80,null,139,null,205,null,436],"deltaPp":4.3693,"id":1,"keywords":["rendering","gaussian","radiance","reconstruction","light"],"label":"3D Gaussian Splatting","labelFull":"3D Gaussian Splatting for Efficient Neural Rendering","papers":413,"papersByYear":[38,null,61,null,93,null,221],"rel":0.8923,"reviewed":true,"share":[4.8969,null,4.4919,null,5.6535,null,9.2662],"size":860,"slope":0.007135},{"claims":[74,null,198,null,191,null,300],"deltaPp":0.3473,"id":2,"keywords":["adversarial","attacks","robustness","perturbations","black-box"],"label":"Adversarial Robustness","labelFull":"Adversarial Attack and Defense Methods for Vision Models","papers":328,"papersByYear":[37,null,87,null,82,null,122],"rel":0.0728,"reviewed":true,"share":[4.768,null,6.4065,null,4.9848,null,5.1153],"size":763,"slope":-0.00019},{"claims":[87,null,152,null,232,null,166],"deltaPp":-1.1233,"id":3,"keywords":["pose","estimation","human","3d","6d"],"label":"3D/6D Pose Estimation","labelFull":"3D Human Pose and 6D Object Pose Estimation","papers":312,"papersByYear":[38,null,77,null,107,null,90],"rel":-0.2294,"reviewed":true,"share":[4.8969,null,5.6701,null,6.5046,null,3.7736],"size":637,"slope":-0.001268},{"claims":[50,null,139,null,216,null,167],"deltaPp":0.1142,"id":4,"keywords":["label","loss","samples","semi-supervised","learning"],"label":"Pseudo-Labeling","labelFull":"Pseudo-Labeling for Semi- and Self-Supervised Learning","papers":363,"papersByYear":[31,null,92,null,142,null,98],"rel":0.0286,"reviewed":true,"share":[3.9948,null,6.7747,null,8.6322,null,4.109],"size":572,"slope":0.0011},{"claims":[48,null,107,null,150,null,212],"deltaPp":2.3322,"id":5,"keywords":["state-of-the-art","benchmark","outperforms","nuscenes","compared"],"label":"State-of-the-Art Performance","labelFull":"State-of-the-Art Results on Standard Vision Benchmarks","papers":485,"papersByYear":[46,null,100,null,142,null,197],"rel":0.3934,"reviewed":true,"share":[5.9278,null,7.3638,null,8.6322,null,8.26],"size":517,"slope":0.004133},{"claims":[94,null,132,null,127,null,159],"deltaPp":-2.2708,"id":6,"keywords":["action","recognition","temporal","activity","egocentric"],"label":"Temporal Action Localization","labelFull":"Temporal Action Localization and Recognition in Videos","papers":240,"papersByYear":[43,null,59,null,60,null,78],"rel":-0.4098,"reviewed":true,"share":[5.5412,null,4.3446,null,3.6474,null,3.2704],"size":512,"slope":-0.003755},{"claims":[114,null,132,null,157,null,107],"deltaPp":-3.5317,"id":7,"keywords":["face","facial","expression","identity","recognition"],"label":"Face Representation Learning","labelFull":"Face Representation Learning for Recognition and Synthesis","papers":210,"papersByYear":[44,null,54,null,61,null,51],"rel":-0.6229,"reviewed":true,"share":[5.6701,null,3.9764,null,3.7082,null,2.1384],"size":510,"slope":-0.005432},{"claims":[66,null,118,null,158,null,138],"deltaPp":-1.291,"id":8,"keywords":["video","frames","temporal","event","representations"],"label":"Efficient Video Representation Learning","labelFull":"Efficient Spatio-Temporal Video Representation Learning","papers":309,"papersByYear":[38,null,81,null,104,null,86],"rel":-0.2636,"reviewed":true,"share":[4.8969,null,5.9647,null,6.3222,null,3.6059],"size":480,"slope":-0.001758},{"claims":[53,null,128,null,161,null,121],"deltaPp":-0.931,"id":9,"keywords":["object detection","detector","few-shot object","proposal","coco"],"label":"Low-Resource Object Detection","labelFull":"Object Detection with Weak, Few-Shot, and Open-Set Supervision","papers":248,"papersByYear":[30,null,67,null,81,null,70],"rel":-0.2408,"reviewed":true,"share":[3.866,null,4.9337,null,4.924,null,2.935],"size":463,"slope":-0.001401},{"claims":[22,null,63,null,91,null,226],"deltaPp":4.162,"id":10,"keywords":["style","editing","content","text-to-image","generation"],"label":"Style-Guided Text-to-Image","labelFull":"Style-Guided Text-to-Image Generation and Editing","papers":213,"papersByYear":[10,null,27,null,46,null,130],"rel":3.2296,"reviewed":true,"share":[1.2887,null,1.9882,null,2.7964,null,5.4507],"size":402,"slope":0.006647},{"claims":[37,null,91,null,92,null,158],"deltaPp":1.1728,"id":11,"keywords":["driving","trajectory","lane","vehicle","autonomous"],"label":"Driving Trajectory Prediction","labelFull":"Trajectory Prediction and Planning for Autonomous Driving","papers":161,"papersByYear":[14,null,37,null,39,null,71],"rel":0.6501,"reviewed":true,"share":[1.8041,null,2.7246,null,2.3708,null,2.9769],"size":378,"slope":0.001582},{"claims":[35,null,108,null,99,null,90],"deltaPp":0.1184,"id":12,"keywords":["domain","adaptation","source","target","unsupervised"],"label":"Unsupervised Domain Adaptation","labelFull":"Unsupervised Domain Adaptation Methods and Benchmarks","papers":175,"papersByYear":[16,null,54,null,53,null,52],"rel":0.0574,"reviewed":true,"share":[2.0619,null,3.9764,null,3.2219,null,2.1803],"size":332,"slope":-0.0002},{"claims":[58,null,76,null,102,null,56],"deltaPp":-2.3084,"id":13,"keywords":["depth","monocular","estimation","rgb","stereo"],"label":"Monocular Depth Estimation","labelFull":"Deep Learning Approaches for Monocular Depth Estimation","papers":154,"papersByYear":[28,null,44,null,51,null,31],"rel":-0.6398,"reviewed":true,"share":[3.6082,null,3.2401,null,3.1003,null,1.2998],"size":292,"slope":-0.003533},{"claims":[0,null,0,null,2,null,261],"deltaPp":7.5052,"id":14,"keywords":["diffusion","text-to-image","latent","pre-trained","motion"],"label":"Diffusion Model Control","labelFull":"Control and Conditioning of Text-to-Image & Video Diffusion","papers":180,"papersByYear":[0,null,0,null,1,null,179],"rel":null,"reviewed":true,"share":[0.0,null,0.0,null,0.0608,null,7.5052],"size":263,"slope":0.011288},{"claims":[77,null,52,null,68,null,62],"deltaPp":-2.3893,"id":15,"keywords":["tracking","trackers","multi-object","siamese","association"],"label":"Visual Object Tracking","labelFull":"Visual Object Tracking: Benchmarks, Datasets, Real-Time Methods","papers":107,"papersByYear":[27,null,24,null,30,null,26],"rel":-0.6867,"reviewed":true,"share":[3.4794,null,1.7673,null,1.8237,null,1.0901],"size":259,"slope":-0.003556},{"claims":[16,null,58,null,91,null,71],"deltaPp":0.3465,"id":16,"keywords":["point","cloud","registration","completion","3d"],"label":"3D Point Cloud Learning","labelFull":"Deep Learning for 3D Point Cloud Processing","papers":138,"papersByYear":[10,null,34,null,55,null,39],"rel":0.2689,"reviewed":true,"share":[1.2887,null,2.5037,null,3.3435,null,1.6352],"size":236,"slope":0.00094},{"claims":[4,null,50,null,76,null,90],"deltaPp":1.9226,"id":17,"keywords":["distillation","knowledge","student","teacher","loss"],"label":"Knowledge Distillation","labelFull":"Knowledge Distillation Techniques for Efficient Model Training","papers":123,"papersByYear":[2,null,26,null,43,null,52],"rel":7.4606,"reviewed":true,"share":[0.2577,null,1.9146,null,2.614,null,2.1803],"size":220,"slope":0.003234},{"claims":[13,null,61,null,78,null,67],"deltaPp":0.6819,"id":18,"keywords":["few-shot","zero-shot","classification","learning","class-incremental"],"label":"Few-Shot Image Classification","labelFull":"Few-Shot Classification and Generalized Zero-Shot Learning","papers":130,"papersByYear":[10,null,32,null,41,null,47],"rel":0.5291,"reviewed":true,"share":[1.2887,null,2.3564,null,2.4924,null,1.9706],"size":219,"slope":0.001091},{"claims":[16,null,49,null,54,null,92],"deltaPp":0.8947,"id":19,"keywords":["surface","meshes","signed","distance","implicit"],"label":"Neural Implicit Surface Reconstruction","labelFull":"Neural Implicit Surface Reconstruction Via SDFs","papers":122,"papersByYear":[9,null,29,null,35,null,49],"rel":0.7714,"reviewed":true,"share":[1.1598,null,2.1355,null,2.1277,null,2.0545],"size":211,"slope":0.001338},{"claims":[23,null,26,null,44,null,115],"deltaPp":1.0982,"id":20,"keywords":["motion","human","sequences","prediction","control"],"label":"Human Motion Generation","labelFull":"Human Motion Generation, Prediction, and Control","papers":113,"papersByYear":[11,null,18,null,24,null,60],"rel":0.7747,"reviewed":true,"share":[1.4175,null,1.3255,null,1.459,null,2.5157],"size":208,"slope":0.001714},{"claims":[30,null,49,null,60,null,64],"deltaPp":0.1042,"id":21,"keywords":["audio-visual","sound","speaker","separation","embeddings"],"label":"Audio-Visual Segmentation","labelFull":"Audio-Visual Segmentation, Parsing, and Source Separation","papers":68,"papersByYear":[7,null,11,null,26,null,24],"rel":0.1155,"reviewed":true,"share":[0.9021,null,0.81,null,1.5805,null,1.0063],"size":203,"slope":0.000542},{"claims":[38,null,47,null,74,null,34],"deltaPp":-1.9157,"id":22,"keywords":["flow","optical","scene","motion","estimation"],"label":"Optical Flow Estimation","labelFull":"Deep Learning Methods for Optical Flow Estimation","papers":112,"papersByYear":[23,null,27,null,37,null,25],"rel":-0.6463,"reviewed":true,"share":[2.9639,null,1.9882,null,2.2492,null,1.0482],"size":193,"slope":-0.002743},{"claims":[0,null,1,null,10,null,180],"deltaPp":4.7379,"id":23,"keywords":["diffusion","sampling","stable","conditional","latent"],"label":"Efficient Diffusion Model Adaptation","labelFull":"Efficient Fine-Tuning, Merging, and Sampling for Diffusion Models","papers":118,"papersByYear":[0,null,1,null,4,null,113],"rel":null,"reviewed":true,"share":[0.0,null,0.0736,null,0.2432,null,4.7379],"size":191,"slope":0.007192},{"claims":[28,null,42,null,48,null,73],"deltaPp":-0.3643,"id":24,"keywords":["pruning","latency","memory","computational","accuracy"],"label":"Efficient Vision Models","labelFull":"Efficient Real-Time Computer Vision Models","papers":160,"papersByYear":[23,null,34,null,41,null,62],"rel":-0.1229,"reviewed":true,"share":[2.9639,null,2.5037,null,2.4924,null,2.5996],"size":191,"slope":-0.000552},{"claims":[6,null,20,null,31,null,130],"deltaPp":2.9196,"id":25,"keywords":["vision-language","visual","text","alignment","encoder"],"label":"Efficient Vision-Language Alignment","labelFull":"Efficient Alignment and Compression for Vision-Language Models","papers":125,"papersByYear":[5,null,14,null,21,null,85],"rel":4.5314,"reviewed":true,"share":[0.6443,null,1.0309,null,1.2766,null,3.5639],"size":187,"slope":0.004502},{"claims":[15,null,50,null,37,null,83],"deltaPp":0.9397,"id":26,"keywords":["navigation","reinforcement learning","policy","planning","exploration"],"label":"Vision-and-Language Navigation","labelFull":"Reinforcement Learning for Vision-and-Language Navigation","papers":103,"papersByYear":[8,null,26,null,22,null,47],"rel":0.9115,"reviewed":true,"share":[1.0309,null,1.9146,null,1.3374,null,1.9706],"size":185,"slope":0.001121},{"claims":[30,null,48,null,26,null,70],"deltaPp":-0.0401,"id":27,"keywords":["captioning","image","grounding","visual","generated"],"label":"Grounded Image Captioning","labelFull":"Grounded and Controllable Image Captioning","papers":80,"papersByYear":[13,null,14,null,14,null,39],"rel":-0.0239,"reviewed":true,"share":[1.6753,null,1.0309,null,0.8511,null,1.6352],"size":174,"slope":-0.00015},{"claims":[4,null,27,null,44,null,92],"deltaPp":1.039,"id":28,"keywords":["anomaly","detection","defect","outlier","normal"],"label":"Visual Anomaly Detection","labelFull":"Visual Anomaly Detection for Industrial Inspection and Video","papers":63,"papersByYear":[3,null,12,null,14,null,34],"rel":2.6875,"reviewed":true,"share":[0.3866,null,0.8837,null,0.8511,null,1.4256],"size":167,"slope":0.001542},{"claims":[45,null,66,null,29,null,17],"deltaPp":-2.0292,"id":29,"keywords":["person","re-identification","visible-infrared","unsupervised","domain"],"label":"Person Re-Identification","labelFull":"Person Re-Identification: Cross-Domain, Unsupervised, Video","papers":76,"papersByYear":[19,null,30,null,17,null,10],"rel":-0.8288,"reviewed":true,"share":[2.4485,null,2.2091,null,1.0334,null,0.4193],"size":157,"slope":-0.003632},{"claims":[32,null,36,null,48,null,40],"deltaPp":-0.624,"id":30,"keywords":["quantization","post-training","quantized","compression","error"],"label":"Post-Training Quantization","labelFull":"Post-Training Quantization for Low-Bit Neural Networks","papers":74,"papersByYear":[12,null,16,null,24,null,22],"rel":-0.4035,"reviewed":true,"share":[1.5464,null,1.1782,null,1.459,null,0.9224],"size":156,"slope":-0.000796},{"claims":[4,null,39,null,60,null,52],"deltaPp":0.9132,"id":31,"keywords":["3d object","object detection","monocular 3d","kitti","waymo"],"label":"Image-based 3D Object Detection","labelFull":"Image-based 3D Object Detection on Driving Datasets","papers":97,"papersByYear":[3,null,26,null,37,null,31],"rel":2.3621,"reviewed":true,"share":[0.3866,null,1.9146,null,2.2492,null,1.2998],"size":155,"slope":0.001537},{"claims":[26,null,37,null,51,null,38],"deltaPp":-1.7092,"id":32,"keywords":["imagenet","top-1","accuracy","resnet","classification"],"label":"ImageNet Classification","labelFull":"ImageNet Classification: Accuracy, Efficiency, Scaling","papers":135,"papersByYear":[24,null,33,null,45,null,33],"rel":-0.5526,"reviewed":true,"share":[3.0928,null,2.43,null,2.7356,null,1.3836],"size":152,"slope":-0.002411},{"claims":[31,null,26,null,26,null,60],"deltaPp":-0.1659,"id":33,"keywords":["question","answering","vqa","visual","reasoning"],"label":"Knowledge-Based VQA","labelFull":"Knowledge-Based Visual Question Answering","papers":74,"papersByYear":[13,null,13,null,12,null,36],"rel":-0.099,"reviewed":true,"share":[1.6753,null,0.9573,null,0.7295,null,1.5094],"size":143,"slope":-0.000363},{"claims":[48,null,39,null,31,null,23],"deltaPp":-3.8426,"id":34,"keywords":["convolutional","activation","networks","layer","cnn"],"label":"Efficient CNNs","labelFull":"Efficient CNN Architectures and Lightweight Modules","papers":116,"papersByYear":[36,null,32,null,29,null,19],"rel":-0.8283,"reviewed":true,"share":[4.6392,null,2.3564,null,1.7629,null,0.7966],"size":141,"slope":-0.006061},{"claims":[9,null,25,null,46,null,57],"deltaPp":0.7813,"id":35,"keywords":["compression","jpeg","psnr","rate-distortion","entropy"],"label":"Learned Image & Video Compression","labelFull":"Learned Image and Video Compression","papers":76,"papersByYear":[5,null,15,null,22,null,34],"rel":1.2126,"reviewed":true,"share":[0.6443,null,1.1046,null,1.3374,null,1.4256],"size":137,"slope":0.001288},{"claims":[34,null,29,null,40,null,34],"deltaPp":-1.3584,"id":36,"keywords":["optimization","programming","solution","relaxation","linear"],"label":"Efficient Optimization Solvers","labelFull":"Efficient Optimization Solvers for Computer Vision","papers":87,"papersByYear":[19,null,21,null,21,null,26],"rel":-0.5548,"reviewed":true,"share":[2.4485,null,1.5464,null,1.2766,null,1.0901],"size":137,"slope":-0.002173},{"claims":[30,null,27,null,57,null,18],"deltaPp":-0.8663,"id":37,"keywords":["text","spotting","scene","recognition","textspotter"],"label":"Scene Text Spotting","labelFull":"Scene Text Spotting: Joint Detection and Recognition","papers":49,"papersByYear":[9,null,13,null,20,null,7],"rel":-0.7469,"reviewed":true,"share":[1.1598,null,0.9573,null,1.2158,null,0.2935],"size":132,"slope":-0.00117},{"claims":[8,null,20,null,50,null,49],"deltaPp":0.4878,"id":38,"keywords":["continual","forgetting","incremental","learning","catastrophic"],"label":"Continual Learning","labelFull":"Continual Learning with Forgetting Mitigation in Vision","papers":74,"papersByYear":[5,null,16,null,26,null,27],"rel":0.7571,"reviewed":true,"share":[0.6443,null,1.1782,null,1.5805,null,1.1321],"size":127,"slope":0.000933},{"claims":[24,null,34,null,33,null,34],"deltaPp":-0.537,"id":39,"keywords":["camera","correspondences","rolling","ransac","calibration"],"label":"Minimal Solvers for Camera Calibration","labelFull":"Minimal Solvers and RANSAC for Camera Calibration","papers":65,"papersByYear":[11,null,18,null,15,null,21],"rel":-0.3788,"reviewed":true,"share":[1.4175,null,1.3255,null,0.9119,null,0.8805],"size":125,"slope":-0.001012},{"claims":[19,null,31,null,44,null,25],"deltaPp":-0.3632,"id":40,"keywords":["localization","supervised object","weakly supervised","cross-view","location"],"label":"Weakly Supervised Object Localization","labelFull":"Weakly Supervised Object Localization in Images","papers":71,"papersByYear":[9,null,18,null,25,null,19],"rel":-0.3132,"reviewed":true,"share":[1.1598,null,1.3255,null,1.5198,null,0.7966],"size":119,"slope":-0.000448},{"claims":[15,null,31,null,47,null,26],"deltaPp":-0.5728,"id":41,"keywords":["video object","object segmentation","video instance","instance segmentation","memory"],"label":"Video Object Segmentation","labelFull":"Video Object Segmentation and Instance Segmentation","papers":67,"papersByYear":[9,null,17,null,27,null,14],"rel":-0.4939,"reviewed":true,"share":[1.1598,null,1.2518,null,1.6413,null,0.587],"size":119,"slope":-0.000664},{"claims":[3,null,20,null,27,null,64],"deltaPp":0.9551,"id":42,"keywords":["restoration","degradation","image","all-in-one","real-world"],"label":"All-in-One Image Restoration","labelFull":"All-in-One Image Restoration for Real-World Degradations","papers":59,"papersByYear":[3,null,11,null,13,null,32],"rel":2.4705,"reviewed":true,"share":[0.3866,null,0.81,null,0.7903,null,1.3417],"size":114,"slope":0.001423},{"claims":[13,null,18,null,28,null,53],"deltaPp":0.3589,"id":43,"keywords":["inpainting","image","guided","patterns","masks"],"label":"Image Inpainting","labelFull":"Image And Video Inpainting With Mask And Pattern Guidance","papers":55,"papersByYear":[6,null,8,null,14,null,27],"rel":0.4642,"reviewed":true,"share":[0.7732,null,0.5891,null,0.8511,null,1.1321],"size":112,"slope":0.000669},{"claims":[4,null,26,null,31,null,51],"deltaPp":1.4132,"id":44,"keywords":["inference","time","speed","efficiency","computation"],"label":"Efficient Inference","labelFull":"Efficient Inference for Vision Models","papers":104,"papersByYear":[4,null,25,null,29,null,46],"rel":2.7414,"reviewed":true,"share":[0.5155,null,1.8409,null,1.7629,null,1.9287],"size":112,"slope":0.002081},{"claims":[6,null,9,null,25,null,70],"deltaPp":1.4551,"id":45,"keywords":["modality","multimodal","missing","includes","alignment"],"label":"Modality-Agnostic Multimodal Learning","labelFull":"Modality-Agnostic Multimodal Learning for Missing Modality","papers":80,"papersByYear":[4,null,9,null,20,null,47],"rel":2.8227,"reviewed":true,"share":[0.5155,null,0.6627,null,1.2158,null,1.9706],"size":110,"slope":0.002459},{"claims":[23,null,35,null,18,null,34],"deltaPp":-0.1893,"id":46,"keywords":["hand","pose","interacting","contact","3d"],"label":"3D Hand Pose Estimation","labelFull":"3D Hand Pose Estimation with Interacting Hands and Objects","papers":45,"papersByYear":[7,null,11,null,10,null,17],"rel":-0.2098,"reviewed":true,"share":[0.9021,null,0.81,null,0.6079,null,0.7128],"size":110,"slope":-0.000385},{"claims":[28,null,23,null,23,null,34],"deltaPp":-1.568,"id":47,"keywords":["retrieval","hashing","cross-modal","image-text","query"],"label":"Image Retrieval","labelFull":"Image Retrieval and Cross-Modal Image-Text Retrieval","papers":72,"papersByYear":[19,null,16,null,16,null,21],"rel":-0.6404,"reviewed":true,"share":[2.4485,null,1.1782,null,0.9726,null,0.8805],"size":108,"slope":-0.002455},{"claims":[14,null,31,null,24,null,34],"deltaPp":-0.0604,"id":48,"keywords":["clothing","garment","body","mesh","human"],"label":"3D Clothed Human Reconstruction","labelFull":"3D Reconstruction of Clothed Humans from Images and Video","papers":46,"papersByYear":[6,null,12,null,11,null,17],"rel":-0.0781,"reviewed":true,"share":[0.7732,null,0.8837,null,0.6687,null,0.7128],"size":103,"slope":-0.000198},{"claims":[39,null,34,null,21,null,8],"deltaPp":-1.3787,"id":49,"keywords":["saliency","salient","maps","prediction","detection"],"label":"Visual Saliency Detection","labelFull":"Visual Saliency Detection and Prediction for Images & Video","papers":40,"papersByYear":[12,null,14,null,10,null,4],"rel":-0.8916,"reviewed":true,"share":[1.5464,null,1.0309,null,0.6079,null,0.1677],"size":102,"slope":-0.00228},{"claims":[7,null,30,null,38,null,27],"deltaPp":0.1104,"id":50,"keywords":["super-resolution","reference-based","low-resolution","image","efficient"],"label":"Image Super-Resolution","labelFull":"Image Super-Resolution: Efficient and Reference-Based Methods","papers":63,"papersByYear":[5,null,18,null,22,null,18],"rel":0.1713,"reviewed":true,"share":[0.6443,null,1.3255,null,1.3374,null,0.7547],"size":102,"slope":0.000172},{"claims":[0,null,1,null,7,null,93],"deltaPp":2.6834,"id":51,"keywords":["prompt","text","visual","tuning","learning"],"label":"Multimodal Prompt Learning","labelFull":"Multimodal Prompt Learning for Vision-Language Models","papers":68,"papersByYear":[0,null,1,null,3,null,64],"rel":null,"reviewed":true,"share":[0.0,null,0.0736,null,0.1824,null,2.6834],"size":101,"slope":0.00408},{"claims":[15,null,23,null,38,null,25],"deltaPp":-0.7917,"id":52,"keywords":["image generation","stylegan","image synthesis","stylegan2","latent"],"label":"GAN Image Synthesis","labelFull":"GANs for High-Fidelity Image Synthesis","papers":77,"papersByYear":[12,null,17,null,30,null,18],"rel":-0.512,"reviewed":true,"share":[1.5464,null,1.2518,null,1.8237,null,0.7547],"size":101,"slope":-0.000902},{"claims":[33,null,14,null,13,null,39],"deltaPp":-0.8724,"id":53,"keywords":["gaze","eye","estimation","targets","egocentric"],"label":"Gaze Estimation and Following","labelFull":"Gaze Estimation and Following for Visual Attention","papers":38,"papersByYear":[11,null,6,null,8,null,13],"rel":-0.6154,"reviewed":true,"share":[1.4175,null,0.4418,null,0.4863,null,0.5451],"size":99,"slope":-0.001286},{"claims":[14,null,29,null,22,null,34],"deltaPp":0.0234,"id":54,"keywords":["scene graph","graph generation","relationship","state-of-the-art scene","visual"],"label":"Scene Graph Generation","labelFull":"Scene Graph Generation for Visual Relationship Detection","papers":53,"papersByYear":[6,null,14,null,14,null,19],"rel":0.0303,"reviewed":true,"share":[0.7732,null,1.0309,null,0.8511,null,0.7966],"size":99,"slope":-5.5e-05},{"claims":[4,null,36,null,17,null,41],"deltaPp":0.3712,"id":55,"keywords":["layout","room","floorplan","plan","design"],"label":"Room Layout Estimation","labelFull":"Room Layout Estimation and Floorplan Reconstruction","papers":39,"papersByYear":[2,null,14,null,8,null,15],"rel":1.4404,"reviewed":true,"share":[0.2577,null,1.0309,null,0.4863,null,0.6289],"size":98,"slope":0.000284},{"claims":[7,null,12,null,43,null,36],"deltaPp":0.4458,"id":56,"keywords":["color","image quality","quality assessment","quantitative","produces"],"label":"Image Colorization","labelFull":"Neural Image Colorization and Quality Assessment","papers":64,"papersByYear":[5,null,11,null,22,null,26],"rel":0.6919,"reviewed":true,"share":[0.6443,null,0.81,null,1.3374,null,1.0901],"size":98,"slope":0.000932},{"claims":[23,null,17,null,28,null,30],"deltaPp":-1.1425,"id":57,"keywords":["attention","mechanism","recurrent","gated","receptive field"],"label":"Efficient Attention","labelFull":"Efficient Attention Modules for Vision Transformers","papers":80,"papersByYear":[17,null,14,null,24,null,25],"rel":-0.5215,"reviewed":true,"share":[2.1907,null,1.0309,null,1.459,null,1.0482],"size":98,"slope":-0.0015},{"claims":[22,null,35,null,26,null,14],"deltaPp":-0.9563,"id":58,"keywords":["translation","image-to-image","unpaired","unsupervised","mapping"],"label":"Unpaired Image-to-Image Translation","labelFull":"Image-to-Image Translation with Unpaired Data","papers":60,"papersByYear":[11,null,23,null,15,null,11],"rel":-0.6746,"reviewed":true,"share":[1.4175,null,1.6937,null,0.9119,null,0.4612],"size":97,"slope":-0.001825},{"claims":[19,null,31,null,25,null,17],"deltaPp":-0.6147,"id":59,"keywords":["graph","matching","node","edge","incremental"],"label":"Deep Graph Matching","labelFull":"Deep Graph Matching and Incremental Graph Matching","papers":51,"papersByYear":[9,null,15,null,14,null,13],"rel":-0.53,"reviewed":true,"share":[1.1598,null,1.1046,null,0.8511,null,0.5451],"size":92,"slope":-0.001049},{"claims":[10,null,10,null,21,null,51],"deltaPp":1.2036,"id":60,"keywords":["view synthesis","morphing","state-of-the-art view","multiview","synthesis quality"],"label":"Novel View Synthesis","labelFull":"Few-Shot Novel View Synthesis from Sparse Multi-View Images","papers":64,"papersByYear":[4,null,6,null,13,null,41],"rel":2.3348,"reviewed":true,"share":[0.5155,null,0.4418,null,0.7903,null,1.7191],"size":92,"slope":0.00198},{"claims":[4,null,14,null,20,null,52],"deltaPp":0.8744,"id":61,"keywords":["lidar","point","3d object","clouds","semantic segmentation"],"label":"LiDAR Point Cloud Perception","labelFull":"LiDAR Point Cloud Detection and Semantic Segmentation","papers":50,"papersByYear":[2,null,8,null,13,null,27],"rel":3.3931,"reviewed":true,"share":[0.2577,null,0.5891,null,0.7903,null,1.1321],"size":90,"slope":0.001412},{"claims":[11,null,31,null,29,null,19],"deltaPp":-0.4051,"id":62,"keywords":["cifar-10","cifar-100","imagenet","svhn","cifar100"],"label":"Image Classification Benchmarks","labelFull":"Image Classification Datasets (CIFAR, SVHN, ImageNet)","papers":85,"papersByYear":[9,null,29,null,29,null,18],"rel":-0.3493,"reviewed":true,"share":[1.1598,null,2.1355,null,1.7629,null,0.7547],"size":90,"slope":-0.000794},{"claims":[15,null,27,null,25,null,22],"deltaPp":-0.3151,"id":63,"keywords":["shape","completion","correspondences","3d","implicit"],"label":"Implicit Correspondence Shape Completion","labelFull":"Implicit Correspondence Learning for 3D Shape Completion","papers":61,"papersByYear":[7,null,23,null,17,null,14],"rel":-0.3493,"reviewed":true,"share":[0.9021,null,1.6937,null,1.0334,null,0.587],"size":89,"slope":-0.000803},{"claims":[25,null,14,null,25,null,21],"deltaPp":-0.6535,"id":64,"keywords":["sketch","image retrieval","fine-grained","synthesis","scene"],"label":"Fine-Grained Sketch Retrieval","labelFull":"Fine-Grained Sketch-Based Image Retrieval and Synthesis","papers":29,"papersByYear":[8,null,5,null,7,null,9],"rel":-0.6339,"reviewed":true,"share":[1.0309,null,0.3682,null,0.4255,null,0.3774],"size":85,"slope":-0.000952},{"claims":[6,null,11,null,18,null,50],"deltaPp":0.5685,"id":65,"keywords":["uncertainty","calibration","mitigation","bias","augmentations"],"label":"Confidence Calibration","labelFull":"Uncertainty Calibration and Confidence Estimation","papers":62,"papersByYear":[6,null,8,null,16,null,32],"rel":0.7353,"reviewed":true,"share":[0.7732,null,0.5891,null,0.9726,null,1.3417],"size":85,"slope":0.001045},{"claims":[4,null,47,null,20,null,12],"deltaPp":0.0777,"id":66,"keywords":["architecture search","neural architecture","search space","binary","reinforcement"],"label":"Efficient Neural Architecture Search","labelFull":"Efficient Neural Architecture Search Techniques","papers":47,"papersByYear":[2,null,23,null,14,null,8],"rel":0.3015,"reviewed":true,"share":[0.2577,null,1.6937,null,0.8511,null,0.3354],"size":83,"slope":-0.000305},{"claims":[22,null,10,null,8,null,42],"deltaPp":-0.5309,"id":67,"keywords":["slam","odometry","visual-inertial","rgbd","direct"],"label":"Learning-Based Visual SLAM","labelFull":"Learning-Based Visual SLAM and Visual-Inertial Odometry","papers":37,"papersByYear":[9,null,7,null,6,null,15],"rel":-0.4578,"reviewed":true,"share":[1.1598,null,0.5155,null,0.3647,null,0.6289],"size":82,"slope":-0.000872},{"claims":[12,null,22,null,23,null,23],"deltaPp":-0.1412,"id":68,"keywords":["anti-spoofing","face","forgery","frequency","identity"],"label":"Face Anti-Spoofing","labelFull":"Face Anti-Spoofing: Cross-Dataset Generalization","papers":35,"papersByYear":[5,null,8,null,10,null,12],"rel":-0.2192,"reviewed":true,"share":[0.6443,null,0.5891,null,0.6079,null,0.5031],"size":80,"slope":-0.000202},{"claims":[18,null,25,null,23,null,14],"deltaPp":-0.7793,"id":69,"keywords":["gan","generator","discriminator","latent","conditional"],"label":"GAN Stabilization","labelFull":"Stabilizing and Regularizing GAN Training","papers":42,"papersByYear":[8,null,13,null,15,null,6],"rel":-0.7559,"reviewed":true,"share":[1.0309,null,0.9573,null,0.9119,null,0.2516],"size":80,"slope":-0.001192},{"claims":[3,null,19,null,25,null,28],"deltaPp":0.6228,"id":70,"keywords":["teacher","student","knowledge","distilled","mean"],"label":"Knowledge Distillation","labelFull":"Teacher-Student Knowledge Distillation for Model Compression","papers":52,"papersByYear":[2,null,13,null,16,null,21],"rel":2.4168,"reviewed":true,"share":[0.2577,null,0.9573,null,0.9726,null,0.8805],"size":75,"slope":0.000942},{"claims":[0,null,8,null,15,null,49],"deltaPp":1.0063,"id":71,"keywords":["avatars","head","animation","codec","modular"],"label":"3D Head Avatar Generation","labelFull":"High-Fidelity 3D Human Head Avatar Generation","papers":35,"papersByYear":[0,null,2,null,9,null,24],"rel":null,"reviewed":true,"share":[0.0,null,0.1473,null,0.5471,null,1.0063],"size":72,"slope":0.001709},{"claims":[0,null,7,null,19,null,44],"deltaPp":0.5451,"id":72,"keywords":["grasp","contact","hand-object","6-dof","policy"],"label":"Language-Conditioned 6-DoF Grasping","labelFull":"Language-Conditioned 6-DoF Grasp Detection and Generation","papers":26,"papersByYear":[0,null,4,null,9,null,13],"rel":null,"reviewed":true,"share":[0.0,null,0.2946,null,0.5471,null,0.5451],"size":70,"slope":0.000944},{"claims":[0,null,0,null,49,null,20],"deltaPp":0.6289,"id":73,"keywords":["vision","transformers","locality","vit","attention"],"label":"Efficient Vision Transformers","labelFull":"Efficient Vision Transformer Architectures and Training","papers":47,"papersByYear":[0,null,0,null,32,null,15],"rel":null,"reviewed":true,"share":[0.0,null,0.0,null,1.9453,null,0.6289],"size":69,"slope":0.001916},{"claims":[8,null,15,null,23,null,20],"deltaPp":-0.0543,"id":74,"keywords":["deblurring","blur","defocus","real-world","low-light"],"label":"Real-World Deblurring","labelFull":"Real-World Image and Video Deblurring Datasets and Methods","papers":36,"papersByYear":[4,null,9,null,12,null,11],"rel":-0.1053,"reviewed":true,"share":[0.5155,null,0.6627,null,0.7295,null,0.4612],"size":66,"slope":-4.8e-05},{"claims":[9,null,16,null,16,null,24],"deltaPp":-0.2701,"id":75,"keywords":["interaction","human-object","affordance","body parts","graph"],"label":"Human-Object Interaction Detection","labelFull":"Human-Object Interaction Detection with Graphs & Synthetic Data","papers":34,"papersByYear":[6,null,8,null,8,null,12],"rel":-0.3493,"reviewed":true,"share":[0.7732,null,0.5891,null,0.4863,null,0.5031],"size":65,"slope":-0.000457},{"claims":[1,null,16,null,21,null,27],"deltaPp":0.5839,"id":76,"keywords":["interpolation","video frame","shutter","event","motion"],"label":"Video Frame Interpolation","labelFull":"Video Frame Interpolation Algorithms, Datasets, and Metrics","papers":36,"papersByYear":[1,null,9,null,9,null,17],"rel":4.5299,"reviewed":true,"share":[0.1289,null,0.6627,null,0.5471,null,0.7128],"size":65,"slope":0.000818},{"claims":[15,null,17,null,24,null,8],"deltaPp":-1.0821,"id":77,"keywords":["gradient","updating","deep","neural networks","regularization"],"label":"Optimizer Design","labelFull":"Optimizer Design and Regularization for Deep Neural Networks","papers":48,"papersByYear":[11,null,11,null,18,null,8],"rel":-0.7634,"reviewed":true,"share":[1.4175,null,0.81,null,1.0942,null,0.3354],"size":64,"slope":-0.001481},{"claims":[5,null,6,null,18,null,34],"deltaPp":0.4132,"id":78,"keywords":["emotion","talking","lip","head","speech"],"label":"Audio-Driven Talking Face Generation","labelFull":"Audio-Driven Talking Face Generation With Emotional Control","papers":31,"papersByYear":[2,null,4,null,9,null,16],"rel":1.6034,"reviewed":true,"share":[0.2577,null,0.2946,null,0.5471,null,0.6709],"size":63,"slope":0.000746},{"claims":[9,null,16,null,7,null,30],"deltaPp":-0.0604,"id":79,"keywords":["denoising","noise","burst","noisy","image"],"label":"Deep Image Denoising","labelFull":"Deep Learning for Image Denoising","papers":37,"papersByYear":[6,null,8,null,6,null,17],"rel":-0.0781,"reviewed":true,"share":[0.7732,null,0.5891,null,0.3647,null,0.7128],"size":62,"slope":-0.000203},{"claims":[5,null,25,null,14,null,18],"deltaPp":0.1104,"id":80,"keywords":["attention","spatial","pyramid","spatiotemporal","channel"],"label":"Spatial-Channel Attention","labelFull":"Spatial-Channel Attention Mechanisms for Vision Transformers","papers":49,"papersByYear":[5,null,15,null,11,null,18],"rel":0.1713,"reviewed":true,"share":[0.6443,null,1.1046,null,0.6687,null,0.7547],"size":62,"slope":-5.2e-05},{"claims":[0,null,0,null,21,null,39],"deltaPp":1.0063,"id":81,"keywords":["token","masked","visual","vision","modeling"],"label":"Vision Token Pruning","labelFull":"Token Pruning and Compression for Vision Transformers","papers":42,"papersByYear":[0,null,0,null,18,null,24],"rel":null,"reviewed":true,"share":[0.0,null,0.0,null,1.0942,null,1.0063],"size":60,"slope":0.002057},{"claims":[19,null,20,null,9,null,11],"deltaPp":-1.2979,"id":82,"keywords":["convolution","group","kernels","input","spatial"],"label":"Adaptive Convolution","labelFull":"Adaptive Convolutional Kernels and Spatially Varying Filters","papers":40,"papersByYear":[13,null,10,null,8,null,9],"rel":-0.7747,"reviewed":true,"share":[1.6753,null,0.7364,null,0.4863,null,0.3774],"size":59,"slope":-0.002072},{"claims":[9,null,26,null,6,null,15],"deltaPp":-0.4378,"id":83,"keywords":["crowd","counting","density","learning mechanism","scale"],"label":"Crowd Counting","labelFull":"Deep Learning for Crowd Counting and Localization","papers":26,"papersByYear":[6,null,9,null,3,null,8],"rel":-0.5662,"reviewed":true,"share":[0.7732,null,0.6627,null,0.1824,null,0.3354],"size":56,"slope":-0.000897},{"claims":[5,null,17,null,8,null,26],"deltaPp":0.1973,"id":84,"keywords":["generative","energy-based","latent","variational","rules"],"label":"Latent Energy-Based Models","labelFull":"Latent Energy-Based Models for Generative Modeling","papers":43,"papersByYear":[4,null,15,null,7,null,17],"rel":0.3827,"reviewed":true,"share":[0.5155,null,1.1046,null,0.4255,null,0.7128],"size":56,"slope":-4.4e-05},{"claims":[6,null,10,null,14,null,26],"deltaPp":0.1973,"id":85,"keywords":["multi-task","task-specific","learning","correlations","shared"],"label":"Efficient Multi-Task Learning","labelFull":"Efficient Multi-Task Learning with Shared Representations","papers":38,"papersByYear":[4,null,6,null,11,null,17],"rel":0.3827,"reviewed":true,"share":[0.5155,null,0.4418,null,0.6687,null,0.7128],"size":56,"slope":0.000409},{"claims":[6,null,9,null,16,null,23],"deltaPp":-0.0962,"id":86,"keywords":["keypoints","refine","localization","score","components"],"label":"Semantic Keypoint Discovery","labelFull":"Semantic Keypoint Discovery and Localization","papers":30,"papersByYear":[4,null,7,null,9,null,10],"rel":-0.1866,"reviewed":true,"share":[0.5155,null,0.5155,null,0.5471,null,0.4193],"size":54,"slope":-0.000128},{"claims":[0,null,0,null,19,null,35],"deltaPp":0.9644,"id":87,"keywords":["token","selection","masked","attention","clustering"],"label":"Token Pruning","labelFull":"Token Pruning for Efficient Vision Transformers","papers":39,"papersByYear":[0,null,0,null,16,null,23],"rel":null,"reviewed":true,"share":[0.0,null,0.0,null,0.9726,null,0.9644],"size":54,"slope":0.001933},{"claims":[7,null,10,null,16,null,20],"deltaPp":0.1134,"id":88,"keywords":["matching","correspondence","map","vectorized","template"],"label":"Semantic Correspondence","labelFull":"Learning Dense Semantic Correspondence for Visual Matching","papers":37,"papersByYear":[4,null,5,null,13,null,15],"rel":0.22,"reviewed":true,"share":[0.5155,null,0.3682,null,0.7903,null,0.6289],"size":53,"slope":0.000381},{"claims":[4,null,10,null,14,null,25],"deltaPp":0.5327,"id":89,"keywords":["qualitative","quantitative","disentanglement","demonstrate","theoretical"],"label":"Qualitative & Quantitative Evaluation","labelFull":"Qualitative and Quantitative Evaluation of Vision Methods","papers":52,"papersByYear":[4,null,9,null,14,null,25],"rel":1.0334,"reviewed":true,"share":[0.5155,null,0.6627,null,0.8511,null,1.0482],"size":53,"slope":0.000893},{"claims":[11,null,19,null,13,null,9],"deltaPp":-0.5636,"id":90,"keywords":["relation","reasoning","relationship","social","entity"],"label":"Scene Graph Generation","labelFull":"Scene Graph Generation with Graph and Transformer Models","papers":31,"papersByYear":[6,null,12,null,8,null,5],"rel":-0.7289,"reviewed":true,"share":[0.7732,null,0.8837,null,0.4863,null,0.2096],"size":52,"slope":-0.001044},{"claims":[8,null,5,null,9,null,30],"deltaPp":-0.1443,"id":91,"keywords":["counterfactual","explanations","causal","decisions","multimodal"],"label":"Counterfactual Explanations","labelFull":"Counterfactual and Causal Explanations for Model Decisions","papers":31,"papersByYear":[6,null,5,null,5,null,15],"rel":-0.1866,"reviewed":true,"share":[0.7732,null,0.3682,null,0.304,null,0.6289],"size":52,"slope":-0.000249},{"claims":[11,null,10,null,15,null,14],"deltaPp":-0.4858,"id":92,"keywords":["feature","clustering","pixel","image","classification"],"label":"Clustering-Based Image Classification","labelFull":"Clustering-Based Feature Learning for Image Classification","papers":42,"papersByYear":[8,null,8,null,13,null,13],"rel":-0.4712,"reviewed":true,"share":[1.0309,null,0.5891,null,0.7903,null,0.5451],"size":50,"slope":-0.000628},{"claims":[1,null,3,null,1,null,45],"deltaPp":1.0032,"id":93,"keywords":["multimodal","prompting","instruction","gpt-4v","reasoning"],"label":"Multimodal LLM Safety","labelFull":"Safety, Alignment, and Prompting for Multimodal LLMs","papers":32,"papersByYear":[1,null,3,null,1,null,27],"rel":7.7828,"reviewed":true,"share":[0.1289,null,0.2209,null,0.0608,null,1.1321],"size":50,"slope":0.001425},{"claims":[6,null,6,null,21,null,16],"deltaPp":-0.0512,"id":94,"keywords":["blind","deconvolution","blur","kernel","deblurring"],"label":"Blind Image Deblurring","labelFull":"Blind Image Deblurring and Deconvolution with Deep Learning","papers":21,"papersByYear":[3,null,3,null,7,null,8],"rel":-0.1324,"reviewed":true,"share":[0.3866,null,0.2209,null,0.4255,null,0.3354],"size":49,"slope":2.6e-05},{"claims":[17,null,9,null,10,null,12],"deltaPp":-0.6505,"id":95,"keywords":["stereo","disparity","matching","confidence estimation","refinement"],"label":"Stereo Matching Networks","labelFull":"Deep Stereo Matching Networks for Disparity Estimation","papers":27,"papersByYear":[7,null,7,null,7,null,6],"rel":-0.7211,"reviewed":true,"share":[0.9021,null,0.5155,null,0.4255,null,0.2516],"size":48,"slope":-0.001021},{"claims":[5,null,7,null,9,null,27],"deltaPp":0.3262,"id":96,"keywords":["understanding","video","grounding","temporal","alignment"],"label":"Temporal Video-Language Grounding","labelFull":"Temporal Video-Language Grounding and Understanding","papers":33,"papersByYear":[3,null,4,null,9,null,17],"rel":0.8438,"reviewed":true,"share":[0.3866,null,0.2946,null,0.5471,null,0.7128],"size":48,"slope":0.000616},{"claims":[0,null,5,null,16,null,25],"deltaPp":0.3774,"id":97,"keywords":["spiking","neural networks","timesteps","asynchronous","vanilla"],"label":"Efficient Spiking Neural Networks","labelFull":"Efficient Training and Inference for Spiking Neural Networks","papers":17,"papersByYear":[0,null,2,null,6,null,9],"rel":null,"reviewed":true,"share":[0.0,null,0.1473,null,0.3647,null,0.3774],"size":46,"slope":0.000675},{"claims":[1,null,10,null,17,null,18],"deltaPp":0.542,"id":98,"keywords":["query","support","prototypes","set","matching"],"label":"Query-Based Object Detection","labelFull":"Query-Based Detection and Segmentation with Transformers","papers":41,"papersByYear":[1,null,9,null,15,null,16],"rel":4.2048,"reviewed":true,"share":[0.1289,null,0.6627,null,0.9119,null,0.6709],"size":46,"slope":0.000938},{"claims":[0,null,1,null,2,null,43],"deltaPp":1.2998,"id":99,"keywords":["text-to-3d","generation","shape","4d","state-of-the-art 3d"],"label":"Text-to-3D Generation","labelFull":"Diffusion-Based Text-to-3D Generation","papers":34,"papersByYear":[0,null,1,null,2,null,31],"rel":null,"reviewed":true,"share":[0.0,null,0.0736,null,0.1216,null,1.2998],"size":46,"slope":0.001974},{"claims":[7,null,11,null,14,null,13],"deltaPp":-0.312,"id":100,"keywords":["real images","annotated","condition","synthetic","single-image"],"label":"Synthetic-to-Real Transfer","labelFull":"Synthetic-to-Real Transfer Learning With Few Real Images","papers":38,"papersByYear":[6,null,11,null,10,null,11],"rel":-0.4035,"reviewed":true,"share":[0.7732,null,0.81,null,0.6079,null,0.4612],"size":45,"slope":-0.000569}],"totals":{"claims":[3718,0,6628,0,8132,0,11741],"papers":[776,0,1358,0,1645,0,2385]},"years":[2018,2019,2020,2021,2022,2023,2024]}