Guillaume Chiron

Detecting and tracking honeybees in 3D at the beehive entrance using stereo vision

In response to recent needs of biologists, we lay the foundations for a real-time stereo vision-based system for monitoring flying honeybees in three dimensions at the beehive entrance. Tracking bees is a challenging task as they are numerous, small, and fast-moving targets with chaotic motion. Contrary to current state-of-the-art approaches, we propose to tackle the problem in 3D space. We present a stereo vision-based system that is able to detect bees at the beehive entrance and is sufficiently reliable for tracking. Furthermore, we propose a detect-before-track approach that employs two innovating methods: hybrid segmentation using both intensity and depth images, and tuned 3D multi-target tracking based on the Kalman filter and Global Nearest Neighbor. Tests on robust ground truths for segmentation and tracking have shown that our segmentation and tracking methods clearly outperform standard 2D approaches.

3D Tracking of Honeybees Enhanced by Environmental Context

This paper summarizes an approach based on stereo vision to recover honeybee trajectories in 3D at the beehive entrance. The 3D advantage offered by stereo vision is crucial to overcome the rough constraints of the application (number of bees, target dynamics and light). Biologists have highlighted the close scale influence of the environment on bees dynamics. We propose to transpose this idea to enhance our tracking process based on Global Nearest Neighbors. Our method normalizes track/observation association costs that are originally not uniformly distributed over the scene. Therefore, the structure of the scene is needed in order to compute relative distances with the targets. The beehive and especially the flight board is the referent environment for bees, so we propose a method to reconstruct the flight board surface from the noisy and incomplete disparity maps provided by the stereo camera.

Impact of OCR errors on the use of digital libraries: towards a better access to information

Digital collections are increasingly used for a variety of purposes. In Europe only, we can conservatively estimate that tens of thousands of users consult digital libraries daily. The usages are often motivated by qualitative and quantitative research. However, caution must be advised as most digitized documents are indexed through their OCRed version, which is far from perfect, especially for ancient documents. In this paper, we aim to estimate the impact of OCR errors on the use of a major online platform: The Gallica digital library from the National Library of France. It accounts for more than 100M OCRed documents and receives 80M search queries every year. In this context, we introduce two main contributions. First, an original corpus of OCRed documents composed of 12M characters along with the corresponding gold standard is presented and provided, with an equal share of English- and French-written documents. Next, statistics on OCR errors have been computed thanks to a novel alignment method introduced in this paper. Making use of all the user queries submitted to the Gallica portal over 4 months, we take advantage of our error model to propose an indicator for predicting the relative risk that queried terms mismatch targeted resources due to OCR errors, underlining the critical extent to which OCR quality impacts on digital library access.

Hybrid Image Retrieval in Digital Libraries

While digital heritage libraries historically took advantage of OCR to index their printed collections, the access to iconographic resources has not progressed in the same way, and the latter remain in the shadows. Today, it would be possible to make better use of these resources, especially by leveraging the illustrations recognized thanks to the OCR produced during the last two decades. This work presents an ETL (extract-transform-load) approach to this need, that aims to: Identify iconography wherever it may be found; Enrich the illustrations metadata with deep learning approaches; Load it all into a web app for hybrid image retrieval.

A Multi-agent Model for Image Browsing and Retrieval

This paper presents a new and original model for image browsing and retrieval based on a reactive multi-agent system oriented toward visualization and user interaction. Each agent represents an image. This model simplifies the problem of mapping a high-dimensional feature space onto a 2D screen interface and allows intuitive user interaction. Within a unify and local model, as opposed to global traditional CBIR, we present how agents can interact through an attraction/repulsion model. These forces are computed based on the visual and textual similarities between an agent and its neighbors. This unique model allows to do several tasks, like image browsing and retrieval, single/multiple querying, performing relevance feedback with positive/nagative examples, all with heteregeneous data (image visual content and text keywords). Specific adjustments are proposed to allow this model to work with large image databases. Preliminary results on two image databases show the feasability of this model compared with traditional CBIR.