SOTAVerified

Image Matting

Image Matting is the process of accurately estimating the foreground object in images and videos. It is a very important technique in image and video editing applications, particularly in film production for creating visual effects. In case of image segmentation, we segment the image into foreground and background by labeling the pixels. Image segmentation generates a binary image, in which a pixel either belongs to foreground or background. However, Image Matting is different from the image segmentation, wherein some pixels may belong to foreground as well as background, such pixels are called partial or mixed pixels. In order to fully separate the foreground from the background in an image, accurate estimation of the alpha values for partial or mixed pixels is necessary.

Source: Automatic Trimap Generation for Image Matting

Image Source: Real-Time High-Resolution Background Matting

Papers

Showing 1–25 of 225 papers

TitleStatusHype
ViTMatte: Boosting Image Matting with Pretrained Plain Vision TransformersCode4
ZIM: Zero-Shot Image Matting for AnythingCode3
What Matters When Repurposing Diffusion Models for General Dense Perception Tasks?Code3
Robust High-Resolution Video Matting with Temporal GuidanceCode3
Real-Time High-Resolution Background MattingCode3
Alfie: Democratising RGBA Image Generation With No $Code2
RenderMe-360: A Large Digital Asset Library and Benchmarks Towards High-fidelity Head AvatarsCode2
Robust Human Matting via Semantic GuidanceCode2
Memory Efficient Matting with Adaptive Token RoutingCode2
MatteFormer: Transformer-Based Image Matting via Prior-TokensCode2
Referring Image MattingCode2
PyMatting: A Python Library for Alpha MattingCode2
MODNet: Real-Time Trimap-Free Portrait Matting via Objective DecompositionCode2
Towards Natural Image Matting in the Wild via Real-Scenario PriorCode2
Deep PCB To COCO ConvertorCode2
Fast Multi-Level Foreground EstimationCode2
Background Matting: The World is Your Green ScreenCode2
In-Context MattingCode2
MaSS13K: A Matting-level Semantic Segmentation BenchmarkCode2
Matte Anything: Interactive Natural Image Matting with Segment Anything ModelsCode2
Castle in the Sky: Dynamic Sky Replacement and Harmonization in VideosCode2
Matting AnythingCode2
MaGGIe: Masked Guided Gradual Human Instance MattingCode2
Deep Image Matting: A Comprehensive SurveyCode2
Adaptive Human Matting for Dynamic VideosCode1
Show:102550
← PrevPage 1 of 9Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1DIMMSE14—Unverified
2IndexNet-MattingMSE13—Unverified
3Context-Aware MattingMSE8.2—Unverified
4SIMMSE5.8—Unverified
5LSAMattingMSE5.4—Unverified
6FBAMattingMSE5.3—Unverified
7PP-MattingMSE5—Unverified
8LFPNetMSE4.1—Unverified
9MatteFormerMSE4—Unverified
10TMFNetMSE3.6—Unverified
#ModelMetricClaimedVerifiedStatus
1SHMCSAD61.5—Unverified
2LFSAD36.12—Unverified
3HATTSAD28.01—Unverified
4SHMSAD17.81—Unverified
5GFM(r)SAD10.89—Unverified
6GFM(d)SAD10.26—Unverified
7GFM(r2b)SAD10.24—Unverified
8GFM(r')SAD9.66—Unverified
9StyleMatteSAD9.6—Unverified
#ModelMetricClaimedVerifiedStatus
1LFSAD42.95—Unverified
2HATTSAD25.99—Unverified
3SHMSAD21.56—Unverified
4GFMSAD13.2—Unverified
5P3M-Net (r)SAD8.73—Unverified
6StyleMatteSAD6.97—Unverified
7P3M-Net (v)SAD6.24—Unverified
#ModelMetricClaimedVerifiedStatus
1LFSAD191.74—Unverified
2SHMSAD170.44—Unverified
3U2NETSAD83.46—Unverified
4GFMSAD52.66—Unverified
5AIM-NetSAD43.92—Unverified
6DiffMatteSAD16.31—Unverified
#ModelMetricClaimedVerifiedStatus
1CAMMSE4.5—Unverified
2BMMSE1.33—Unverified
3IMMSE1.16—Unverified
4Adobe LS-GANMSE0.97—Unverified
#ModelMetricClaimedVerifiedStatus
1PP-MattingSAD40.69—Unverified
2DCAMSAD31.27—Unverified
3ViTMatteSAD17.05—Unverified
4DiffMatteSAD15.5—Unverified
#ModelMetricClaimedVerifiedStatus
1MODNet+MAD0.81—Unverified
#ModelMetricClaimedVerifiedStatus
1MODNet+ (Our)MAD0.97—Unverified