We devolop identity recognition technology based gait and deploy it on CCTV cameras.
Limitations of identification conditions.
Long distance, accestories, easily affected by external environment (light, camera quality).
Differences in movement and body shape are key to making a person's gait unique
We proposed new method for Gait Recognition, called Gait CAL.
This method using Autoencoders + BiLSTM and Gait Energy Image for Gait Encoding.
Compared to other methods, our method improves performance and accuracy significantly
Evaluate on CASIA-B dataset
Graduation project score: 9.5
Draft for ICIP2023
created with
HTML editor .