Project overview
Autism Spectrum Disorder (ASD) is a neurodevelopmental condition that can affect sensory, motor, speech and memory-related functions. Identifying early behavioural indicators is challenging, making the study of measurable neurobehavioural patterns important. This research introduces a Blended Multi-scale Attention Network (BMANet) to classify different levels of ASD-like characteristics in an experimental model. The proposed model combines Residual Dilated Spatial Pyramid Pooling (RDSPP) with spatial and channel-wise attention modules to identify important behavioural features at different scales. It analyses features such as locomotion and anxiety-related behaviour to improve classification.