Images

BSD (Berkeley Segmentation Dataset)

Introduced by David R. Martin et al. in A Database of Human Segmented Natural Images and its Application to Evaluating Segmentation Algorithms and Measuring Ecological Statistics

BSD is a dataset used frequently for image denoising and super-resolution. Of the subdatasets, BSD100 is aclassical image dataset having 100 test images proposed by Martin et al.. The dataset is composed of a large variety of images ranging from natural images to object-specific such as plants, people, food etc. BSD100 is the testing set of the Berkeley segmentation dataset BSD300.

Source: A Deep Journey into Super-resolution: A Survey

Homepage

Benchmarks

Add a new result Link an existing benchmark

Task	Dataset Variant	Best Model
Image Super-Resolution	BSD100 - 4x upscaling	DRCT-L
Image Super-Resolution	BSD100 - 2x upscaling	WaveMixSR-V2
Image Super-Resolution	BSD100 - 3x upscaling	Hi-IR-L
Grayscale Image Denoising	BSD68 sigma15	ADL
Grayscale Image Denoising	BSD68 sigma25	KBNet
Grayscale Image Denoising	BSD68 sigma50	ADL
Image Super-Resolution	BSD100 - 8x upscaling	DRLN+
Density Estimation	BSDS300	TAN
Color Image Denoising	BSD68 sigma25	RIDNet
Color Image Denoising	BSD68 sigma15	RIDNet
Grayscale Image Denoising	BSD200 sigma70	RC-Net
Blind Super-Resolution	BSD100 - 2x upscaling	DCLS
Grayscale Image Denoising	BSD200 sigma50	RC-Net
Blind Super-Resolution	BSD100 - 4x upscaling	DCLS
Grayscale Image Denoising	BSD200 sigma30	RC-Net
Grayscale Image Denoising	BSD68 sigma70	N3Net
Salt-And-Pepper Noise Removal	BSD300 Noise Level 50%	CNN
Grayscale Image Denoising	BSD68 sigma35	FFDNet
Grayscale Image Denoising	BSD68 sigma75	FFDNet
Image Super-Resolution	BSDS100 - 2x upscaling	DBPN-RES-MR64-3
Salt-And-Pepper Noise Removal	BSD300 Noise Level 30%	CNN
Grayscale Image Denoising	BSD200 sigma10	RC-Net
Salt-And-Pepper Noise Removal	BSD300 Noise Level 70%	CNN
Blind Super-Resolution	BSD100 - 3x upscaling	DCLS
Grayscale Image Denoising	BSD68 sigma10	Residual Dense Network +
Grayscale Image Denoising	BSD68 sigma30	Residual Dense Network +
Unified Image Restoration	BSD68 sigma25	DA-RCOT
Color Image Denoising	BSD68 sigma10	Residual Dense Network +
Color Image Denoising	BSD68 sigma30	Residual Dense Network +
Color Image Denoising	BSD68 sigma70	Residual Dense Network +
Grayscale Image Denoising	BSD68 sigma5	BUIFD75
Grayscale Image Denoising	BSD68 sigma20	BUIFD75
Grayscale Image Denoising	BSD68 sigma40	BUIFD75
Grayscale Image Denoising	BSD68 sigma45	BUIFD75
Grayscale Image Denoising	BSD68 sigma55	BUIFD75
Grayscale Image Denoising	BSD68 sigma60	BUIFD75
Grayscale Image Denoising	BSD68 sigma65	BUIFD75
Color Image Denoising	BSD68 sigma5	Deep CNN Denoiser
Color Image Denoising	BSD68 sigma35	Deep CNN Denoiser
Color Image Denoising	BSD68 sigma75	CSCNet
Image Super-Resolution	BSDS100 - 4x upscaling	DBPN-RES-MR64-3
Image Super-Resolution	BSDS100 - 8x upscaling	DBPN-RES-MR64-3
Image Super-Resolution	BSD100 - 16x upscaling	ABPN
Image Super-Resolution	BSD200 - 2x upscaling	CSRCNN
Compressive Sensing	BSDS100 - 2x upscaling	AMPA-Net
Compressive Sensing	BSD68 CS=50%	AMPA-Net
Image Denoising	BSD68 sigma30	R3L