A5049656720- Advanced Image Processing Techniques
- Image Enhancement Techniques
- AI in cancer detection
- Machine Learning and ELM
- Advanced Neural Network Applications
- Image and Signal Denoising Methods
- Process Optimization and Integration
- Neural Networks and Applications
- Digital Imaging for Blood Diseases
Tokyo Institute of Technology
2024
Real-world images prevalently contain different varieties of degradation, such as motion blur and luminance noise. Computer vision recognition models trained on clean perform poorly degraded images. Previously, several works have explored how to image classification while training a single model for each degradation. Nevertheless, it becomes challenging host degradation limited hardware applications estimate parameters correctly at the run-time. This work proposes method effectively...
Image classification is extensively used in various applications such as satellite imagery, autonomous driving, smartphones, and healthcare. Most of the images to train models can be considered ideal, i.e., without any degradation either due corruption pixels camera sensors, sudden shake blur, or compression a specific format. In this paper, we have proposed novel CNN-based architecture for image degraded based on intermediate layer knowledge distillation data augmentation approach cutout...
Image classification is a typical computer vision task widely used in practical applications. The images for training image networks are often clean, i.e., without any degradation. However, Convolutional neural trained on clean perform poorly degraded or corrupted the real world. In this study, we effectively utilize robust data augmentation (DA) with knowledge distillation to improve performance of images. We first categorize augmentations into geometric-and-color and cut-and-delete DAs....