Computational Baby Learning

TL;DR

This paper introduces a computational model inspired by how babies learn to recognize new objects with minimal supervision, leveraging prior knowledge, exemplar learning, and video context to improve object detection from few samples.

Contribution

The paper presents a novel slightly-supervised object detection framework that mimics baby learning, combining pre-trained CNNs, exemplar learning, and video-based instance discovery to outperform full-data models.

Findings

Effective detection with very few samples per category.

Utilizes unlabeled videos to enhance learning.

Outperforms state-of-the-art full-training methods.

Abstract

Intuitive observations show that a baby may inherently possess the capability of recognizing a new visual concept (e.g., chair, dog) by learning from only very few positive instances taught by parent(s) or others, and this recognition capability can be gradually further improved by exploring and/or interacting with the real instances in the physical world. Inspired by these observations, we propose a computational model for slightly-supervised object detection, based on prior knowledge modelling, exemplar learning and learning with video contexts. The prior knowledge is modeled with a pre-trained Convolutional Neural Network (CNN). When very few instances of a new concept are given, an initial concept detector is built by exemplar learning over the deep features from the pre-trained CNN. Simulating the baby's interaction with physical world, the well-designed tracking solution is then used to discover more diverse instances from the massive online unlabeled videos. Once a positive instance is detected/identified with high score in each video, more variable instances possibly from different view-angles and/or different distances are tracked and accumulated. Then the concept detector can be fine-tuned based on these new instances. This process can be repeated again and again till we obtain a very mature concept detector. Extensive experiments on Pascal VOC-07/10/12 object detection datasets well demonstrate the effectiveness of our framework. It can beat the state-of-the-art full-training based performances by learning from very few samples for each object category, along with about 20,000 unlabeled videos.