Extracting features of body sway for baggage weight classification

We measure the slight body sway of a standing person from depth video recorded by an overhead camera. Frequency analysis of the temporal sway features is used to classify the person's baggage as heavy or light.

Even when people appear to be standing still, their bodies sway slightly. We hypothesize that this body sway changes with the weight of the baggage they are carrying. This study investigates whether these small movements can help classify baggage as heavy or light.

An overhead camera records a person standing while carrying baggage. It captures depth images, which include information about the distance to visible surfaces. We extract the body movements over time and examine how much of the sway consists of slower or faster changes. This process, called frequency analysis, provides information for distinguishing light from heavy baggage.

We collected video recordings and used them to test the method. The results confirmed that information extracted from body sway helped classify baggage as heavy or light. The approach focuses on the carrier’s small movements rather than the visible appearance of the baggage itself.

• 山口 優太，中山 晴貴，神谷 卓也, 西山 正志, 岩井 儀雄, 櫛田 大輔,
  身体動揺の時系列深度画像における時間手掛りを用いた手荷物の軽重認識,
  精密工学会誌, vol. 88, no. 1, pp. 91 - 101, January 2022.
• Yuta Yamaguchi, Takuya Kamitani, Masashi Nishiyama, Yoshio Iwai, Daisuke Kushida,
  Extracting features of body sway for baggage weight classification,
  Proceedings of IEEE 9th Global Conference on Consumer Electronics (GCCE), OS-QOL.6, pp. 443 - 446, October 2020.

• Baggage weight classification using body sway