Image Processing Projects

Abstract: Probabilistic model-based segmentation uses the hidden Markov random field (HMRF) to represent image class labels. Existing HMRF models consider either the number of neighboring pixels with similar class labels or their spatial distance. This…

Abstract: This paper introduces a Retinex-based fractional-order variational model for severely low-light images. The proposed method allows greater control over regularization extent than integer-order regularization methods. To improve estimates, we decompose directly in the image…

Abstract: Most hand-crafted key-point detectors like Harris corner, MSER, SIFT, and SURF use pre-designed structures to detect corners, blobs, and junctions in images. These detectors are inflexible due to pre-designed structures. Non-uniform illumination also greatly…

Abstract: MDC is an efficient source coding method for error-prone transmission over multiple channels. This paper introduces a polyphase down-sampling-based MDC (NPDS-MDC) for image signals. Three steps encode our proposed NPDS-MDC. First, we apply quincunx…

Abstract: Aerial scenes, unlike nature scenes, have many objects crowded on the surface in bird’s view, requiring more discriminative features and local semantics. Most convolution neural networks (ConvNets) used for scene classification depict global semantics…

Abstract: Generalized nucleus segmentation can speed up visual biomarker development and validation for new digital pathology datasets. The MoNuSeg 2018 Challenge developed generalizable nuclei segmentation methods for digital pathology. Read Also: Python Projects 32 teams…

Abstract: Deep learning and computer vision researchers study interpretable data representation. Representation disentanglement is effective for this task, but existing works cannot easily handle problems that require manipulating and recognizing data across multiple domains. We…

Abstract: We propose a theoretical solution to 2D and 3D affine invariant feature matching, a longstanding computer vision problem. Our two-stage Grassmannian Graph (GrassGraph) framework robustly recovers correspondences between two unorganized, affinely related feature (point)…

Abstract: Reversible color-to-grayscale conversion (RCGC) embeds chromatic information into a grayscale image so it can be reconstructed later. It must authenticate an information-embedded image to ensure its integrity. No RCGC algorithm considers this factor. This…

Abstract: DPC-CT is useful for soft-tissue and low-atomic-number sample analysis. Due to implementation constraints, DPC-CT often has incomplete projections. Incomplete data challenges conventional reconstruction algorithms. They involve time-consuming, noisy parameter selection operations. We present DPC-CT,…

Abstract: Environmental, illumination, atmospheric, and temporal changes cause hyperspectral image spectral variability. It may cause significant unmixing estimation errors. Extended linear mixing models generate large-scale nonsmooth ill-posed inverse problems to address this issue. Regularization strategies…

Abstract: In low-visibility scenes, image enhancement is a crucial pre-processing step for many computer vision applications. We develop a unified two-pathway model inspired by biological vision, especially early visual mechanisms, to enhance images using low…