Abstract
Instance segmentation of foods is an important technology to ensure the food success rate of meal-assisting robotics. However, due to foods have strong intraclass variability, interclass similarity, and complex physical properties, which leads to more challenges in recognition, localization, and contour acquisition of foods. To address the above issues, this paper proposed a novel method for instance segmentation of foods. Specifically, in backbone network, deformable convolution was introduced to enhance the ability of YOLOv8 architecture to capture finer-grained spatial information, and efficient multiscale attention based on cross-spatial learning was introduced to improve sensitivity and expressiveness of multiscale inputs. In neck network, classical convolution and C2f modules were replaced by lightweight convolution GSConv and improved VoV-GSCSP aggregation module, respectively, to improve inference speed of models. We abbreviated it as the DEG-YOLOv8n-seg model. The proposed method was compared with baseline model and several state-of-the-art (SOTA) segmentation models on datasets, respectively. The results show that the DEG-YOLOv8n-seg model has higher accuracy, faster speed, and stronger robustness. Specifically, the DEG-YOLOv8n-seg model can achieve 84.6% Box_mAP@0.5 and 84.1% Mask_mAP@0.5 accuracy at 55.2 FPS and 11.1 GFLOPs. The importance of adopting data augmentation and the effectiveness of introducing deformable convolution, EMA, and VoV-GSCSP were verified by ablation experiments. Finally, the DEG-YOLOv8n-seg model was applied to experiments of food instance segmentation for meal-assisting robots. The results show that the DEG-YOLOv8n-seg can achieve better instance segmentation of foods. This work can promote the development of intelligent meal-assisting robotics technology and can provide theoretical foundations for other tasks of the computer vision field with some reference value.
Similar content being viewed by others
Data availability
Data will be made available on request.
References
Daehyung, P., Yuuna, H., Charles, C.: A multimodal anomaly detector for robot-assisted feeding using an LSTM-based variational autoencoder. IEEE Rob. Autom. Lett. 3(3), 1544–1551 (2018)
Jihyeon, H., Sangin, P., Chang-Hwan, I., Laehyun, K.: A hybrid brain–computer interface for real-life food assist robot control. Sensors 21, 4578 (2021)
Nabil, E., Aman, B.: A learning from demonstration framework for implementation of a feeding task. Ency. Semant. Comput. Robot Intell. 2(1), 1850001 (2018)
Tejas, K., Maria, K., Graser, A.: Application of reinforcement learning to a robotic drinking assistant. Robotics 9(1), 1–15 (2019)
Fei, L., Hongliu, Y., Wentao, W., Changcheng, Q.: I-feed: a robotic platform of an assistive feeding robot for the disabled elderly population. Technol. Health Care 2, 1–5 (2020)
Fei, L., Peng, X., Hongliu, Y.: Robot-assisted feeding: a technical application that combines learning from demonstration and visual interaction. Technol. Health Care 1, 1–6 (2020)
Yuhe, F., Lixun, Z., Xingyuan, W., Keyi, W., Lan, W., Zhenhan, W., Feng, X., Jinghui, Z., Chao, W.: Rheological thixotropy and pasting properties of food thickening gums orienting at improving food holding rate. Appl. Rheol. 32, 100–121 (2022)
Yuhe, F., Lixun, Z., Jinghui, Z., Yunqin, Z., Xingyuan, W.: Viscoelasticity and friction of solid foods measurement by simulating meal-assisting robot. Int. J. Food Prop. 25(1), 2301–2319 (2022)
Yuhe, F., Lixun, Z., Canxing, Z., Xingyuan, W., Keyi, W., Jinghui, Z.: Motion behavior of non-Newtonian fluid-solid interaction foods. J. Food Eng. 347, 111448 (2023)
Yuhe, F., Lixun, Z., Canxing, Z., Feng, X., Zhenhan, W., Xingyuan, W., Lan, W.: Contact forces and motion behavior of non-Newtonian fluid–solid food by coupled SPH–FEM method. J. Food Sci. 88(6), 2536–2556 (2023)
Weng, Z., Meng, F., Liu, S., Zhang, Y., Zheng, Z., Gong, C.: Cattle face recognition based on a two-branch convolutional neural network. Comput. Electron. Agric. 196, 106871 (2022)
Jinhai, W., Zongyin, Z., Lufeng, L., Huiling, W., Wei, W., Mingyou, C., Shaoming, L.: DualSeg: Fusing transformer and CNN structure for image segmentation in complex vineyard environment. Comput. Electron. Agr. 206, 107682 (2023)
Chan, Z., Pengfei, C., Jing, P., Xiaofan, Y., Changxin, C., Shuqin, T., Yueju, X.: A mango picking vision algorithm on instance segmentation and key point detection from RGB images in an open orchard. Biosyst. Eng. 206, 32–54 (2021)
Jordi, G., Mar, F., Eduard, G., Jochen, H., JosepRamon, M.: Looking behind occlusions: a study on a modal segmentation for robust on-tree apple fruit size estimation. Comput. Electron. Agr. 209, 107854 (2023)
Dandan, W., Dongjian, H.: Fusion of Mask R-CNN and attention mechanism for instance segmentation of apples under complex background. Comput. Electron. Agr. 196, 106864 (2022)
Ang, W., Juanhua, Z., Taiyong, R.: Detection of apple defect using laser-induced light backscattering imaging and convolutional neural network. Comput. Electr. Eng. 81, 106454 (2020)
Tian, Y., Yang, G., Wang, Z., Li, E., Liang, Z.: Instance segmentation of apple flowers using the improved mask R-CNN model. Biosyst. Eng. 193, 264–278 (2020)
Olarewaju, M.: YOLOv5-LiNet: a lightweight network for fruits instance segmentation. PLoS ONE 18(3), e0282297 (2023)
Rajdeep, K., Rakesh, K., Meenu, G.: Food Image-based diet recommendation framework to overcome PCOS problem in women using deep convolutional neural network. Comput. Electr. Eng. 103, 108298 (2022)
Zhu, L., Li, Z., Li, C., Wu, J., Yue, J.: High performance vegetable classification from images based on Alexnet deep learning model. Int. J. Agr. Biol. Eng. 11(4), 217–223 (2018)
Haozan, L., Guihua, W., Yang, H., Mingnan, L., Pei, Y., Yingxue, X.: MVANet: multi-task guided multi-view attention network for Chinese food recognition. IEEE T. Multimedia 23, 3551–3561 (2021)
Eduardo, A., Bhalaji, N., Beatriz, R., Petia, R.: Bayesian deep learning for semantic segmentation of food images. Comput. Electr. Eng. 103, 108380 (2022)
Liu, Y., Han, Z., Liu, X., Wang, J., Wang, C., Liu, R.: Estimation method and research of fruit glycemic load index based on the fusion SE module faster R-CNN. Comput. Electr. Eng. 109, 108696 (2023)
Tang, Z., Zhou, L., Qi, F., Chen, H.: An improved lightweight and real-time YOLOv5 network for detection of surface defects on indocalamus leaves. J. Real-Time Image Pr. 20(14), 1–14 (2023)
Yuhe, F., Lixun, Z., Canxing, Z., Yunqin, Z., Xingyuan, W., Jinghui, Z.: Real-time and accurate meal detection for meal-assisting robots. J. Food Eng. 371, 111996 (2024)
Lingling, F., Hanyu, Z., Jiaxin, Z., Xianghai, W.: Image classification with an RGB-channel nonsubsampled contourlet transform and a convolutional neural network. Neurocomputing 396, 266–277 (2020)
Yu, F., Xinxing, L., Yinggang, Z., Tianhua, X.: Detection of Atlantic salmon residues based on computer vision. J. Food Eng. 358, 111658 (2023)
Kunshan, Y., Jun, S., Chen, C., Min, X., Xin, Z., Yan, C., Yan, T.: Non-destructive detection of egg qualities based on hyperspectral imaging. J. Food Eng. 325, 111024 (2022)
Li, W., Mao, S., Mahoney, A., Petkovic, S., Coyle, J., Sejdic, E.: Deep learning models for bolus segmentation in videofuoroscopic swallow studies. J. Real-Time Image Pr. 21(18), 1–10 (2024)
Yousong, Z., Xu, Z., Chaoyang, Z., Jinqiao, W., Hanqing, L.: Food det: Detecting foods in refrigerator with supervised transformer network. Neurocomputing 379, 162–171 (2020)
Glenn, J.: Ultralytics YOLOv8. https://github.com/ultralytics/ultralytics (2023)
Redmon, J., Divvala, S., Girshick, R., Farhadi, A.: You only look once: unified, real-time object detection. In: IEEE Conf. Comput. Vis. Pattern. Recognit., pp. 779–788 (2016)
Dai, J., Qi, H., Xiong, Y., Li, Y., Zhang, G., Hu, H., Wei, Y.: Deformable convolutional networks. In: IEEE Int. Conf. Comput. Vis., pp. 764–773 (2017)
Daliang, O., Su, H., Guozhong, Z., Mingzhu, L., Huaiyong, G., Jian, Z., Zhijie, H.: Efficient Multi-Scale Attention Module with Cross-Spatial Learning. In: ICASSP, pp. 1–5 (2023)
Hulin, L., Jun, L., Hanbing, W., Zheng, L., Zhenfei, Z., Qiliang, R.: Slim-neck by GsConv: A better design paradigm of detector architectures for autonomous vehicles. In: IEEE Conf. Comput. Vis. Pattern. Recognit., pp. 1–17 (2022)
Tianhua, L., Meng, S., Qinghai, H., Guanshan, Z., Guoying, S., Xiaoming, D., Sen, L.: Tomato recognition and location algorithm based on improved YOLOv5. Comput. Electron. Agr. 208, 107759 (2023)
Wang, C., Bochkovskiy, A., Liao, H.: YOLOv7: Trainable bag-of-freebiessets new state-of-the-art for real-time object detectors. In: IEEE Conf. Comput. Vis. Pattern. Recognit (2022)
Glenn, J.: YOLOv5 release v6.1, https://github.com/ultralytics/yolov5/releases/tag/v6.1 (2022)
He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: IEEE Conf. Comput. Vis. Pattern. Recognit., pp. 770–778 (2016)
Wenjie, Y., Jiachun, W., Jinlai, Z., Kai, G., Ronghua, D., Zhuo, W., Eksan, F., Dingwen, L.: Deformable convolution and coordinate attention for fast cattle detection. Comput. Electron. Agric. 211, 108006 (2023)
Chilukuri, D., Yi, S., Seong, Y.: A robust object detection system with occlusion handling for mobile devices. Comput. Intell. 38(4), 1338–1364 (2022)
Fang, H., Li, J., Tang, H., Xu, C., Zhu, H., Xiu, Y., Li, Y., Lu, C.: Alphapose: Whole-body regional multi-person pose estimation and tracking in real-time. IEEE Trans. Pattern Anal. Mach. Intell. 45(6), 7157–7173 (2022)
Qibin, H., Daquan, Z., Jiashi, F.: Coordinate attention for efficient mobile network design. In: IEEE Conf. Comput. Vis. Pattern. Recognit., pp. 13708–13717 (2021)
Ryan, F., Youngsun, K., Gilwoo, L., Ethan, K.: Robot-assisted feeding: Generalizing skewering strategies across food items on a realistic plate. In: IEEE Conf. Comput. Vis. Pattern. Recognit. arXiv preprint: arXiv:1906.02350 (2021)
Girshick, R.: Fast r-cnn. In: IEEE Conf. Comput. Vis. Pattern. Recognit., pp. 1440–1448 (2015).
Tsungyi, L., Priya, G., Ross, G., Kaiming, H., Piotr, D.: Focal loss for dense object detection. In: IEEE Conf. Comput. Vis. Pattern. Recognit. arXiv:1708.02002 (2017)
Haoyang, Z., Ying, W., Feras, D., Niko, S.: VarifocalNet: An IoU-aware Dense Object Detector. In: IEEE Conf. Comput. Vis. Pattern. Recognit. arXiv:2008.13367v2 (2021)
Wada, K.: https://github.com/wkentaro/labelme (2020)
Jinlai, Z., Lyujie, C., Bo, O., Binbin, L., Jihong, Z., Yujing, C., Yanmei, M., Danfeng, W.: Pointcutmix: Regularization strategy for point cloud classification. In: IEEE Conf. Comput. Vis. Pattern. Recognit. arXiv:2101.01461 (2022)
Su, D., Kong, H., Qiao, Y., Sukkarieh, S.: Data augmentation for deep learning based semantic segmentation and crop-weed classification in agricultural robotics. Comput. Electron. Agric. 190, 106418 (2021)
Shu, L., Lu, Q., Haifang, Q., Jianping, S., Jiaya, J.: Path aggregation network for instance segmentation. In: IEEE Conf. Comput. Vis. Pattern. Recognit. arXiv:1803.01534v4 (2018)
Chengyang, F., Mykhailo, S., Alexander, C.: RetinaMask: Learning to predict masks improves state-of-the-art single-shot detection for free. In: IEEE Conf. Comput. Vis. Pattern. Recognit. arXiv:1901.03353v1 (2019)
Kaiming, H., Georgia, G., Piotr, D., Ross, G.: Mask R-CNN. In: IEEE Conf. Comput. Vis. Pattern. Recognit. (2018)
Daniel, B., Chong, Z., Fanyi, X., Yongjae, L.: YOLACT real-time instance segmentation. In: IEEE Conf. Comput. Vis. Pattern. Recognit. arXiv:1904.02689v2 (2019)
Acknowledgements
The research work is supported by National Key R&D Program of China under grant 2020YFC2007700 and Fundamental Research Funds for the Central Universities of China under grant 3072022CF0703.
Funding
National Key R&D Program of China, 2020YFC2007700, Fundamental Research Funds for the Central Universities of China, 3072022CF0703.
Author information
Authors and Affiliations
Contributions
Yuhe Fan: analysis, experiments, drafting, and revising. Lixun Zhang: funding, methods, reviewed, and revised. Canxing Zheng: results, collecting images, and datasets. Yunqin Zu: datasets and experiments. Keyi Wang: reviewed and revised. Xingyuan Wang: analysis and theory. All authors agree to be accountable for all aspects of the work.
Corresponding author
Ethics declarations
Conflict of interest
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Fan, Y., Zhang, L., Zheng, C. et al. Real-time and accurate model of instance segmentation of foods. J Real-Time Image Proc 21, 80 (2024). https://doi.org/10.1007/s11554-024-01459-z
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s11554-024-01459-z