Sequential Attend, Infer, Repeat: Generative Modelling of Moving Objects
This addresses the challenge of interpretable video analysis for applications like surveillance, though it is incremental as it builds on prior work.
The paper tackles the problem of unsupervised learning for videos of moving objects by presenting SQAIR, a generative model that discovers, tracks, and generates future frames, overcoming limitations of its predecessor in handling overlapping or occluded objects, as demonstrated on multi-MNIST and real-world CCTV data.
We present Sequential Attend, Infer, Repeat (SQAIR), an interpretable deep generative model for videos of moving objects. It can reliably discover and track objects throughout the sequence of frames, and can also generate future frames conditioning on the current frame, thereby simulating expected motion of objects. This is achieved by explicitly encoding object presence, locations and appearances in the latent variables of the model. SQAIR retains all strengths of its predecessor, Attend, Infer, Repeat (AIR, Eslami et. al., 2016), including learning in an unsupervised manner, and addresses its shortcomings. We use a moving multi-MNIST dataset to show limitations of AIR in detecting overlapping or partially occluded objects, and show how SQAIR overcomes them by leveraging temporal consistency of objects. Finally, we also apply SQAIR to real-world pedestrian CCTV data, where it learns to reliably detect, track and generate walking pedestrians with no supervision.