Susumu Takahashi

Neuroscientist@Doshisha Univ. (Spatial navigation, Hippocampus, Place Cell, Episodic memory; Basal ganglia, Parkinson's disease)

[H.SHIMAZAKI @h_shimazaki]

1/14 🧠 Our paper is out in Science Advances!

S. Amin Moosavi, Sai Sumedh R. Hindupur, Hideaki Shimazaki
Population coding under the scale invariance of high-dimensional noise
Science Advances 12, eadz9632, 2026

https://t.co/qlsxNWHVqy

🎨 Robin Hoshino

[Susumu Takahashi @Tsusumu2]

🎉 New publication from Takahashi Lab in JoVE

We present a video protocol demonstrating how vmTracking improves multi-animal pose estimation accuracy and reduces identity swaps.
https://t.co/taz0r1zDIf

#vmTracking #PoseEstimation

[Susumu Takahashi @Tsusumu2]

vmTracking is out at PLoS Biology! A novel protocol for precise multi‐animal tracking via virtual markers, robust ID in dense groups. PLOS Biol: https://t.co/vjrG6YQ7JO

[さらピン！キョウト｜KBS京都ラジオ @kbs_sarakyo]

月曜日のさらピン！始まります✨
本日は同志社大学脳科学研究科の御園生教授にお越しいただきました。初登場です！

また、3時台には歌手の梅谷心愛さんにお越しいただきますよ♪

#サラピン #KBS京都

[Susumu Takahashi @Tsusumu2]

🆕 New version of our paper "vmTracking: Virtual Markers Overcome Occlusion and Crowding in Multi-Animal Pose Tracking" is now available on bioRxiv! 🚀

🔗https://t.co/TkFBJcbd2e

#behavior #PoseTracking #neuroscience #ecology #bioRxiv

[Susumu Takahashi @Tsusumu2]

脳科学研究科では、５月１１日（土）に入試説明会を開催いたします。https://t.co/if36C6d5cv

[Keisuke Fujii (藤井慶輔) @keisuke_fj]

東進タイムズ2024に名古屋大学情報学部コンピュータサイエンス学科と、私の研究室の記事が掲載されました！
「社会を変えるデータサイエンス」という特集の第1回に選んでいただきました。
名大情報学部CS学科、良いところですよ！（と宣伝しました！）

[Susumu Takahashi @Tsusumu2]

(2/2)

2. Make the tracking result into a video. It's recommended to have two or more keypoints.
Important that at least one keypoint is visible during occlusion.

3. Track the video made with maDLC with #DLC or #SLEAP for single animals.

[Susumu Takahashi @Tsusumu2]

Our vmTracking enhances the tracking accuracy of #DeepLabCut, #maDLC, and #SLEAP with a simple yet effective approach, excelling particularly in challenging scenarios such as occlusion and crowding.
https://t.co/VKEWPsxuWf

[Susumu Takahashi @Tsusumu2]

(1/2)
It's natural to question if this demonstration video is just human-annotated.
For the curious, trying it with your own videos is easy, especially with maDLC processed ones.

1. Track with #maDLC or #SLEAP.

[Susumu Takahashi @Tsusumu2]

If you are having trouble with Multi-Animal Pose Tracking, this vmTracking might make you happy.
Look at the comparison video.
https://t.co/Y0rXbtlR3F… #bioRxiv