IEEE

Learn From Past to Future: Exploiting Self-Training and Curriculum Learning in Remote Sensing Class-Incremental Semantic Segmentation

Abstract: Class-incremental semantic segmentation focuses on updating the segmentation model with only new-class samples. Catastrophic forgetting and background shift are the two prevalent challenges.
IEEE

Optimized Ensemble Learning for Semantic Segmentation of Satellite Imagery Using DeepLabV3+ and UNet With PSO and Cross-Dataset Evaluation

Abstract: Semantic segmentation is critical in remote sensing applications such as urban planning, disaster management, and environmental monitoring. However, segmenting complex satellite images ...
GitHub

Symdex-100

Semantic fingerprints for intent-based Python code search — 50–100x faster index lookups, 10–50x fewer tokens for AI agents. Symdex-100 generates compact, structured metadata ("Cyphers") for every ...
GitHub

Vision model for object detection and segmentation using TENNs State Space Model neural network.

State Space Models (SSMs) are emerging as a practical alternative to transformers, offering similar or better performance with significantly fewer parameters and lower compute requirements. Mamba, the ...