Showing 1–5 of 5 results for author: Haucke, T

Align and Distill: Unifying and Improving Domain Adaptive Object Detection

Authors: Justin Kay, Timm Haucke, Suzanne Stathatos, Siqi Deng, Erik Young, Pietro Perona, Sara Beery, Grant Van Horn

Abstract: Object detectors often perform poorly on data that differs from their training set. Domain adaptive object detection (DAOD) methods have recently demonstrated strong results on addressing this challenge. Unfortunately, we identify systemic benchmarking pitfalls that call past results into question and hamper further progress: (a) Overestimation of performance due to underpowered baselines, (b) Inc… ▽ More Object detectors often perform poorly on data that differs from their training set. Domain adaptive object detection (DAOD) methods have recently demonstrated strong results on addressing this challenge. Unfortunately, we identify systemic benchmarking pitfalls that call past results into question and hamper further progress: (a) Overestimation of performance due to underpowered baselines, (b) Inconsistent implementation practices preventing transparent comparisons of methods, and (c) Lack of generality due to outdated backbones and lack of diversity in benchmarks. We address these problems by introducing: (1) A unified benchmarking and implementation framework, Align and Distill (ALDI), enabling comparison of DAOD methods and supporting future development, (2) A fair and modern training and evaluation protocol for DAOD that addresses benchmarking pitfalls, (3) A new DAOD benchmark dataset, CFC-DAOD, enabling evaluation on diverse real-world data, and (4) A new method, ALDI++, that achieves state-of-the-art results by a large margin. ALDI++ outperforms the previous state-of-the-art by +3.5 AP50 on Cityscapes to Foggy Cityscapes, +5.7 AP50 on Sim10k to Cityscapes (where ours is the only method to outperform a fair baseline), and +0.6 AP50 on CFC Kenai to Channel. Our framework, dataset, and state-of-the-art method offer a critical reset for DAOD and provide a strong foundation for future research. Code and data are available: https://github.com/justinkay/aldi and https://github.com/visipedia/caltech-fish-counting. △ Less

Submitted 25 August, 2024; v1 submitted 18 March, 2024; originally announced March 2024.

Comments: 28 pages, 12 figures

SOCRATES: A Stereo Camera Trap for Monitoring of Biodiversity

Authors: Timm Haucke, Hjalmar S. Kühl, Volker Steinhage

Abstract: The development and application of modern technology is an essential basis for the efficient monitoring of species in natural habitats and landscapes to trace the development of ecosystems, species communities, and populations, and to analyze reasons of changes. For estimating animal abundance using methods such as camera trap distance sampling, spatial information of natural habitats in terms of… ▽ More The development and application of modern technology is an essential basis for the efficient monitoring of species in natural habitats and landscapes to trace the development of ecosystems, species communities, and populations, and to analyze reasons of changes. For estimating animal abundance using methods such as camera trap distance sampling, spatial information of natural habitats in terms of 3D (three-dimensional) measurements is crucial. Additionally, 3D information improves the accuracy of animal detection using camera trapping. This study presents a novel approach to 3D camera trapping featuring highly optimized hardware and software. This approach employs stereo vision to infer 3D information of natural habitats and is designated as StereO CameRA Trap for monitoring of biodivErSity (SOCRATES). A comprehensive evaluation of SOCRATES shows not only a $3.23\%$ improvement in animal detection (bounding box $\text{mAP}_{75}$) but also its superior applicability for estimating animal abundance using camera trap distance sampling. The software and documentation of SOCRATES is provided at https://github.com/timmh/socrates △ Less

Submitted 13 October, 2022; v1 submitted 19 September, 2022; originally announced September 2022.

Journal ref: Sensors 2022, 22(23), 9082

Automated Distance Estimation for Wildlife Camera Trapping

Authors: Peter Johanns, Timm Haucke, Volker Steinhage

Abstract: The ongoing biodiversity crisis calls for accurate estimation of animal density and abundance to identify sources of biodiversity decline and effectiveness of conservation interventions. Camera traps together with abundance estimation methods are often employed for this purpose. The necessary distances between camera and observed animals are traditionally derived in a laborious, fully manual or se… ▽ More The ongoing biodiversity crisis calls for accurate estimation of animal density and abundance to identify sources of biodiversity decline and effectiveness of conservation interventions. Camera traps together with abundance estimation methods are often employed for this purpose. The necessary distances between camera and observed animals are traditionally derived in a laborious, fully manual or semi-automatic process. Both approaches require reference image material, which is both difficult to acquire and not available for existing datasets. We propose a fully automatic approach we call AUtomated DIstance esTimation (AUDIT) to estimate camera-to-animal distances. We leverage existing state-of-the-art relative monocular depth estimation and combine it with a novel alignment procedure to estimate metric distances. AUDIT is fully automated and requires neither the comparison of observations in camera trap imagery with reference images nor capturing of reference image material at all. AUDIT therefore relieves biologists and ecologists from a significant workload. We evaluate AUDIT on a zoo scenario dataset unseen during training where we achieve a mean absolute distance estimation error over all animal instances of only 0.9864 meters and mean relative error (REL) of 0.113. The code and usage instructions are available at https://github.com/PJ-cs/DistanceEstimationTracking △ Less

Submitted 11 May, 2022; v1 submitted 9 February, 2022; originally announced February 2022.

ACM Class: I.4.8

Journal ref: Ecological Informatics, Volume 70, September 2022, 101734

Overcoming the Distance Estimation Bottleneck in Estimating Animal Abundance with Camera Traps

Authors: Timm Haucke, Hjalmar S. Kühl, Jacqueline Hoyer, Volker Steinhage

Abstract: The biodiversity crisis is still accelerating, despite increasing efforts by the international community. Estimating animal abundance is of critical importance to assess, for example, the consequences of land-use change and invasive species on community composition, or the effectiveness of conservation interventions. Various approaches have been developed to estimate abundance of unmarked animal p… ▽ More The biodiversity crisis is still accelerating, despite increasing efforts by the international community. Estimating animal abundance is of critical importance to assess, for example, the consequences of land-use change and invasive species on community composition, or the effectiveness of conservation interventions. Various approaches have been developed to estimate abundance of unmarked animal populations. Whereas these approaches differ in methodological details, they all require the estimation of the effective area surveyed in front of a camera trap. Until now camera-to-animal distance measurements are derived by laborious, manual and subjective estimation methods. To overcome this distance estimation bottleneck, this study proposes an automatized pipeline utilizing monocular depth estimation and depth image calibration methods. We are able to reduce the manual effort required by a factor greater than 21 and provide our system at https://timm.haucke.xyz/publications/distance-estimation-animal-abundance △ Less

Submitted 22 December, 2021; v1 submitted 10 May, 2021; originally announced May 2021.

ACM Class: I.4.9; I.5.4

Journal ref: Ecological Informatics, Volume 68, May 2022, 101536

Exploiting Depth Information for Wildlife Monitoring

Authors: Timm Haucke, Volker Steinhage

Abstract: Camera traps are a proven tool in biology and specifically biodiversity research. However, camera traps including depth estimation are not widely deployed, despite providing valuable context about the scene and facilitating the automation of previously laborious manual ecological methods. In this study, we propose an automated camera trap-based approach to detect and identify animals using depth e… ▽ More Camera traps are a proven tool in biology and specifically biodiversity research. However, camera traps including depth estimation are not widely deployed, despite providing valuable context about the scene and facilitating the automation of previously laborious manual ecological methods. In this study, we propose an automated camera trap-based approach to detect and identify animals using depth estimation. To detect and identify individual animals, we propose a novel method D-Mask R-CNN for the so-called instance segmentation which is a deep learning-based technique to detect and delineate each distinct object of interest appearing in an image or a video clip. An experimental evaluation shows the benefit of the additional depth estimation in terms of improved average precision scores of the animal detection compared to the standard approach that relies just on the image information. This novel approach was also evaluated in terms of a proof-of-concept in a zoo scenario using an RGB-D camera trap. △ Less

Submitted 10 February, 2021; originally announced February 2021.