Corpus ID: 237635201

From images in the wild to video-informed image classification

  title={From images in the wild to video-informed image classification},
  author={Marc B{\"o}hlen and Varun Chandola and Wawan Sujarwo and Raunaq Jain},
  • M. Böhlen, V. Chandola, +1 author Raunaq Jain
  • Published 24 September 2021
  • Computer Science, Engineering
  • ArXiv
Image classifiers work effectively when applied on structured images, yet they often fail when applied on images with very high visual complexity. This paper describes experiments applying state-of-the-art object classifiers toward a unique set of ‘images in the wild’ with high visual complexity collected on the island of Bali. The text describes differences between actual images in the wild and images from Imagenet, and then discusses a novel approach combining informational cues particular to… Expand

Figures and Tables from this paper


Improving Image Classification with Location Context
This work tackles the problem of performing image classification with location context, and explores different ways of encoding and extracting features from the GPS coordinates, and shows how to naturally incorporate these features into a Convolutional Neural Network, the current state-of-the-art for most image classification and recognition problems. Expand
Context R-CNN: Long Term Temporal Context for Per-Camera Object Detection
This paper proposes an attention-based approach that allows the model, Context R-CNN, to index into a long term memory bank constructed on a per-camera basis and aggregate contextual features from other frames to boost object detection performance on the current frame. Expand
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
This work introduces a Region Proposal Network (RPN) that shares full-image convolutional features with the detection network, thus enabling nearly cost-free region proposals and further merge RPN and Fast R-CNN into a single network by sharing their convolutionAL features. Expand
Plant Recognition by Inception Networks with Test-time Class Prior Estimation
Performance improvements were achieved by adjusting the CNN predictions according to the estimated change of the class prior probabilities, replacing network parameters with their running averages, and test-time data augmentation. Expand
Deep Reinforcement Learning of Region Proposal Networks for Object Detection
We propose drl-RPN, a deep reinforcement learning-based visual recognition model consisting of a sequential region proposal network (RPN) and an object detector. In contrast to typical RPNs, whereExpand
Image-based classification of plant genus and family for trained and untrained plant species
The results show that shared visual characters are indeed present at higher taxonomic levels, and enable state-of-the-art classification algorithms to learn accurate visual representations of plant genera and families. Expand
Overview of LifeCLEF Plant Identification Task 2019: diving into Data Deficient Tropical Countries
The resources and assessments of the challenge are presented, the approaches and systems employed by the participating research groups are summarized, and an analysis of the main outcomes are provided. Expand
Overview of LifeCLEF 2018: A Large-Scale Evaluation of Species Identification and Recommendation Algorithms in the Era of AI
The methodology of the conducted evaluations as well as the synthesis of the main results and lessons learned are described to push the boundaries of the state-of-the-art in several research directions. Expand
Machine Learning in Ethnobotany
  • M. Böhlen, W. Sujarwo
  • Computer Science
  • 2020 IEEE International Conference on Systems, Man, and Cybernetics (SMC)
  • 2020
A novel approach to ethnobotany documentation that harnesses machine learning opportunities, specifically for the documentation of traditional ecological knowledge with mobile phones in emerging economies is described. Expand
Media Images and the Social Construction of Reality
Ideally, a media system suitable for a democracy ought to provide its readers with some coherent sense of the broader social forces that affect the conditions of their everyday lives. It is difficultExpand