Deep learning-based image retrieval and classification model with multi-similarity using transformer enabled multi-scale Yolov5 with attention mechanism

Key Points

• The aim is to develop a deep learning-based model that improves image retrieval and classification performance.
• Utilized a transformer-enabled Yolov5 architecture
• Incorporated an attention mechanism
• Employed multi-similarity techniques for improved classification
• Achieved significant accuracy improvements in image retrieval
• Enhanced classification performance across various datasets
• Demonstrated efficiency in processing images with varying scales.