Date

2022-9-02

Author

Ertenli, Can Ufuk

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

109
views

54
downloads

Cite This

There have been numerous significant developments for addressing recognition problems in recent years. One important application area of such developments is recognition on streaming data. Efficient inference is typically critical for streaming data sources, especially for real-time applications such as autonomous driving and robot control. For this purpose, this thesis presents StreamDEQ, a method that infers frame-wise representations on videos with minimal per-frame computation. In contrast to conventional methods where compute time grows at least linearly with the network depth, we aim to update the representations in a continuous manner. For this purpose, we leverage the recently emerging implicit layer models, which infer the representation of an image by solving a fixed-point problem. Our main insight is to leverage the slowly changing nature of videos and use the previous frame representation as an initial condition on each frame. This scheme effectively recycles the recent inference computations and greatly reduces the needed processing time. Through extensive experimental analysis, we show that StreamDEQ is able to recover near-optimal representations in a few frames' time and maintain an up-to-date representation throughout the video duration. Our experiments on video semantic segmentation and video object detection show that StreamDEQ achieves on-par accuracy with the baseline (standard MDEQ) while being more than 3x faster.

Subject Keywords

Implicit layer models, Video analysis and understanding, Video object detection, Video semantic segmentation

URI

https://hdl.handle.net/11511/99465

Collections

Graduate School of Natural and Applied Sciences, Thesis