The Keras has provided high level library functions for deep learning, however, it didn't provide any connectors to the database. As Caffe provided standard database usage for dataset preprocessing,
we also plan to implement this key component, to promise fast data access. Also, we also plan to add advanced hashing algorithms, like isotropic hashing etc. To improve the locality of image storage, make it possible to compress with more efficiency.
[x] Implement LevelDB access functionalities to finish fast image access
[x] Implement LMDB access functionalities to finish fast image access
[x] Implement advanced hashing algorithms to extract hashing features for images
[x] Test the Compression Ratio of Hashing Features
The Keras has provided high level library functions for deep learning, however, it didn't provide any connectors to the database. As Caffe provided standard database usage for dataset preprocessing, we also plan to implement this key component, to promise fast data access. Also, we also plan to add advanced hashing algorithms, like isotropic hashing etc. To improve the locality of image storage, make it possible to compress with more efficiency.
[x] Implement LevelDB access functionalities to finish fast image access
[x] Implement LMDB access functionalities to finish fast image access
[x] Implement advanced hashing algorithms to extract hashing features for images
[x] Test the Compression Ratio of Hashing Features