Accurate detection and tracking of ants in indoor and outdoor environments

Abstract

Monitoring social insects' activity is critical for biologists researching their group mechanism. Manually labelling individual insects in a video is labour-intensive. Automated tracking social insects is particularly challenging: (1) individuals are small and similar in appearance; (2) frequent interactions with each other cause severe and long-term occlusion. We propose a detection and tracking framework for ants by: (1) adopting a two-stage object detection framework using ResNet-50 as backbone and coding the position of regions of interest to locate ants accurately; (2) using the ResNet model to develop the appearance descriptors of ants; (3) constructing long-term appearance sequences and combining them with motion information to achieve online tracking. To validate our method, we build a video database of ant colony captured in both indoor and outdoor scenes. We achieve a state-of-the-art performance of 95.7% mMOTA and 81.1% mMOTP in indoor videos, 81.8% mMOTA and 81.9% mMOTP in outdoor videos. Our method runs 6-10 times faster than existing methods for insect tracking. The datasets and code are made publicly available, we aim to contribute to an automated tracking tool for biologists in relevant domains.

Assessing the effect of complex ground types on ground‐dwelling arthropod movements with video monitoring: Dealing with concealed movements under a layer of plant residues

“…In the case of diurnal species, there are more possibilities for distinguishing individuals from the background. Indeed, recording in visible light results in less noise than infrared light (Semenishchev et al, 2018) and standard cameras under visible light return three values per pixel, providing more possibilities to find colors (not just intensities) that distinguish individuals from the background (Sebastian et al, 2010; Wu et al, 2020 preprint). The standard infrared camera used in this experiment returns only one value per pixel (intensity), suggesting that reflective material is probably the best option for nocturnal species.…”

Artificial intelligence versus natural selection: Using computer vision techniques to classify bees and bee mimics

“…In a 2021 study (Spiesman et al, 2021) (Carney et al, 2022;Minakshi et al, 2020), crop pests (Kasinathan et al, 2021), beetles (Venegas et al, 2021), and hornets (Jeong et al, 2020), as well as ants and their movements (Wu et al, 2020). In the realm of investigating bee mimicry using deep neural networks, we are aware of one recent work in 2019, which looks at M € ullerian mimicry among bumble bees across spatial scales (Ezray et al, 2019).…”

NeuroHSMD: Neuromorphic Hybrid Spiking Motion Detector

“…Animals behaviours may include monitoring of livestock (monitor cows, pigs and diseases detection through the analysis of atypical movements) [28], [29]; gathering understanding about complex colonies of insects (e.g. bees and ants have communication mechanisms that enables them to work together to solve complex problems) [30], [31]; and monitoring wildlife (e.g. track movements of shoals of fish or pods of whales) [32], [33].…”