Summary of Human-like Relational Models For Activity Recognition in Video, by Joseph Chrol-cannon et al.
Human-like Relational Models for Activity Recognition in Video
by Joseph Chrol-Cannon, Andrew Gilbert, Ranko Lazic, Adithya Madhusoodanan, Frank Guerin
First submitted to arxiv on: 12 Jul 2021
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed paper aims to improve video activity recognition by developing a more human-like approach that focuses on extracting critical spatio-temporal relations among objects and hands. Building upon deep neural networks, the method interprets videos in sequential temporal phases and extracts specific relationships between objects and hands, which are then used to train random forest classifiers. The proposed approach is applied to a challenging subset of the something-something dataset, demonstrating improved performance on difficult activities compared to traditional neural network baselines. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper proposes a new way to recognize video activities that’s more like how humans do it. Right now, computers are good at recognizing some things in videos, but they struggle with complex actions that involve multiple objects and their relationships. The proposed approach breaks down the video into smaller parts, focusing on specific interactions between objects and hands. This information is then used to train special types of classifiers called random forest models. The results show that this approach works better than traditional computer vision methods for recognizing difficult activities. |
Keywords
» Artificial intelligence » Activity recognition » Neural network » Random forest
