Learning from the Tangram to Solve Mini Visual Tasks
Not on the map. Why?
Abstract (arXiv)
Current pre-training methods in computer vision focus on natural images in the daily-life context. However, abstract diagrams such as icons and symbols are common and important in the real world. This work is inspired by Tangram, a game that requires replicating an abstract pattern from seven dissected shapes. By recording human experience in solving tangram puzzles, we present the Tangram dataset and show that a pre-trained neural model on the Tangram helps solve some mini visual tasks based on low-resolution vision. Extensive experiments demonstrate that our proposed method generates intelligent solutions for aesthetic tasks such as folding clothes and evaluating room layouts. The pre-trained feature extractor can facilitate the convergence of few-shot learning tasks on human handwriting and improve the accuracy in identifying icons by their contours. The Tangram dataset is available at https://github.com/yizhouzhao/Tangram.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 28 of its 57 references that are in the map. First posted on arXiv 12 Dec 2021.
Builds on
28 of 57 references.
References · 28
- 2015Deep Residual Learning for Image Recognitioncited by 51,809
- 2014Microsoft COCO: Common Objects in Contextcited by 17,498
- 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
- 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,415
- 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,861
- 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,283
- 2017Prototypical Networks for Few-shot Learningcited by 3,001
- 2016Generative Adversarial Imitation Learningcited by 1,576
- 2019Self-supervised Visual Feature Learning with Deep Neural Networks: A Surveycited by 656
- 2018Rethinking ImageNet Pre-trainingcited by 376
- 2019Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAMLcited by 359
- 2018Meta-Transfer Learning for Few-Shot Learningcited by 306
16 more
- 2020MetaSDF: Meta-learning Signed Distance Functionscited by 214
- 2017Using millions of emoji occurrences to learn any-domain representations for detecting sentiment, emotion and sarcasmcited by 117
- 2018Fast and Flexible Indoor Scene Synthesis via Deep Convolutional Generative Modelscited by 105
- 2018Benchmarking Neural Network Robustness to Common Corruptions and Surface Variationscited by 99
- 2019In Defense of Pre-trained ImageNet Architectures for Real-time Semantic Segmentation of Road-driving Imagescited by 99
- 2020House-GAN: Relational Generative Adversarial Networks for Graph-constrained House Layout Generationcited by 86
- 2018Human-centric Indoor Scene Synthesis Using Stochastic Grammarcited by 73
- 2020Leveraging the Feature Distribution in Transfer-based Few-Shot Learningcited by 72
- 2019The Omniglot challenge: a 3-year progress reportcited by 54
- 2016Image Aesthetic Assessment: An Experimental Surveycited by 50
- 2018Assessing four Neural Networks on Handwritten Digit Recognition Dataset (MNIST)cited by 8
- 2020G-SimCLR : Self-Supervised Contrastive Learning with Guided Projection via Pseudo Labellingcited by 7
- 2019Learning icons appearance similaritycited by 6
- 2020Iconify: Converting Photographs into Iconscited by 4
- 2019Understanding the Effects of Pre-Training for Object Detectors via Eigenspectrumcited by 3
- 2019Image Aesthetics Assessment using Multi Channel Convolutional Neural Networkscited by 1
Built on by
Citing papers per year: 2022: 1.
1 paper.
- 2022Geoclidean: Few-Shot Generalization in Euclidean Geometry30 Nov 2022