Superpixels
태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.
- 하위 태스크
- 0
- 평가표
- 0
- 보존된 평가 기록
- 0
Superpixel techniques segment an image into regions based on similarity measures that utilize perceptual features, effectively grouping pixels that appear similar. The motivation behind this approach is to generate regions that provide meaningful descriptions while significantly reducing the data volume compared to using every individual pixel. By decreasing the number of primitives, these techniques reduce redundancy and simplify the complexity of recognition tasks. Superpixels replace the rigid structure of individual pixels with delineated regions that preserve meaningful content in the image, thereby aiding the interpretation of the scene’s structure and simplifying subsequent processing tasks. Generally, superpixel techniques rely on measures that evaluate color similarities and the shapes of regions, incorporating edges or significant changes in intensity to define these regions.
논문
이 태스크와 하위 분야의 관련 논문입니다. 실제 성능 비교는 평가 결과에서 확인하세요.
Predicting Depth, Surface Normals and Semantic Labels with a Common Multi-Scale Convolutional Architecture
Deep Deconvolutional Networks for Scene Parsing
Visual Chunking: A List Prediction Framework for Region-Based Object Detection
A Bottom-Up Approach for Automatic Pancreas Segmentation in Abdominal CT Scans
A Statistical Modeling Approach to Computer-Aided Quantification of Dental Biofilm
A Context-aware Delayed Agglomeration Framework for Electron Microscopy Segmentation
Lazy Random Walks for Superpixel Segmentation
SEEDS: Superpixels Extracted via Energy-Driven Sampling
Machine learning of hierarchical clustering to segment 2D and 3D images
출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.