Anders Jørgensen

Long-Term Occupancy Analysis Using Graph-Based Optimisation in Thermal Imagery

This paper presents a robust occupancy analysis system for thermal imaging. Reliable detection of people is very hard in crowded scenes, due to occlusions and segmentation problems. We therefore propose a framework that optimises the occupancy analysis over long periods by including information on the transition in occupancy, when people enter or leave the monitored area. In stable periods, with no activity close to the borders, people are detected and counted which contributes to a weighted histogram. When activity close to the border is detected, local tracking is applied in order to identify a crossing. After a full sequence, the number of people during all periods are estimated using a probabilistic graph search optimisation. The system is tested on a total of 51,000 frames, captured in sports arenas. The mean error for a 30-minute period containing 3-13 people is 4.44 %, which is a half of the error percentage optained by detection only, and better than the results of comparable work. The framework is also tested on a public available dataset from an outdoor scene, which proves the generality of the method.

Detecting gallbladders in chicken livers using spectral analysis

This paper presents a method for detecting gallbladders attached to chicken livers using spectral imaging. Gallbladders can contaminate good livers, making them unfit for human consumption. A data set consisting of chicken livers with and without gallbladders, has been captured using 33 wavelengths within the visible spectrum. This work shows how to reduce the high number of wavelengths while maintaining a high accuracy. A classification tree has be trained to evaluate if a gallbladder is present and whether it is suitable for automatic removal, which could increase profits for the processing plants. As a preliminary study this shows good results with a classification accuracy of 91.7%.

RGB-D Segmentation of Poultry Entrails

This paper presents an approach for automatic visual inspection of chicken entrails in RGB-D data. The point cloud is first over-segmented into supervoxels based on color, spatial and geometric information. Color, position and texture features are extracted from each of the resulting supervoxels and passed to a Random Forest classifier, which classifies the supervoxels as either belonging to heart, lung, liver or misc. The dataset consists of 150 individual entrails, with 30 of these being reserved for evaluation. Segmentation performance is evaluated on a voxel-by-voxel basis, achieving an average Jaccard index of 61.5 % across the four classes of organs. This is a 5.9 % increase over the 58.1 % achieved with features derived purely from 2D.

Diagnosis of Broiler Livers by Classifying Image Patches

The manual health inspection are becoming the bottleneck at poultry processing plants. We present a computer vision method for automatic diagnosis of broiler livers. The non-rigid livers, of varying shape and sizes, are classified in patches by a convolutional neural network, outputting maps with probabilities of the three most common diseases. A Random Forest classifier combines the maps to a single diagnosis. The method classifies 77.6% livers correctly in a problem that is far from trivial.

Prototype Software for Automated Structural Analysis of Systems

Abstract In this paper we present a prototype software tool that is developed to analyse the structural model of automated systems in order to identify redundant information that is hence utilized for Fault detection and Isolation (FDI) purposes. The dedicated algorithms in this software tool use a tri-partite graph that represents the structural model of the system. A component-based approach has been used to address issues such as system complexity and reconfigurability possibilities.

Organ Segmentation in Poultry Viscera Using RGB-D

We present a pattern recognition framework for semantic segmentation of visual structures, that is, multi-class labelling at pixel level, and apply it to the task of segmenting organs in the eviscerated viscera from slaughtered poultry in RGB-D images. This is a step towards replacing the current strenuous manual inspection at poultry processing plants. Features are extracted from feature maps such as activation maps from a convolutional neural network (CNN). A random forest classifier assigns class probabilities, which are further refined by utilizing context in a conditional random field. The presented method is compatible with both 2D and 3D features, which allows us to explore the value of adding 3D and CNN-derived features. The dataset consists of 604 RGB-D images showing 151 unique sets of eviscerated viscera from four different perspectives. A mean Jaccard index of 78.11% is achieved across the four classes of organs by using features derived from 2D, 3D and a CNN, compared to 74.28% using only basic 2D image features.

Automatic Analysis of Activities in Sports Arenas Using Thermal Cameras

The demand for automatically gathered data is a societal trend quickly extending to all aspects of human life. Knowledge on the utilization of public facilities is of interest for optimising use and cutting expenses for the owners. Manual observations are both cumbersome and expensive, and they have a risk of incorrect results due to subjective opinions or lack of interest in the given task. In this paper we present the main results of a 5-year long research project revolving around the real-world application of automatic analysis of activities in sports arenas. Three topics are explored: Counting people, recognising activities, and estimating energy expenditure. The project is based on thermal image data, to preserve privacy while capturing video in public sports arenas. This paper aim to provide an overview of our published methods and results within these three topics and add a discussion of the results and perspectives of this research.