Indoor Scene Recognition Using ARM-based MobileNets Architectures

The rapid development of science and technology has improved the quality of people life. In recent years, the use of microcontrollers has increased due to the rise of edge computing. Based on low cost, low power consumption, and high stability, the controller can be widely used in various fields. In this research, an ARM-based platform is applied with a camera module to perform image recognition tasks. The MQTT protocol is realized to transmit the image recognition results. The MobileNets models are developed with X-CUBE-AI tool to perform transfer learning on indoor scene datasets. The verification results are obtained by training MobileNetV1 and MobileNetV2 structures. The proposed image system indeed achieves the average accuracies of 67.2% and 71.6% for MobileNetV1 and MobileNetV2, respectively.

[1]  Lisong Ou,et al.  Identification Algorithm of Diseased Leaves based on MobileNet Model , 2022, 2022 4th International Conference on Communications, Information System and Computer Engineering (CISCE).

[2]  Venkatesh,et al.  Fine-tuned MobileNet Classifier for Classification of Strawberry and Cherry Fruit Types , 2021, 2021 International Conference on Computer Communication and Informatics (ICCCI).

[3]  Mark Sandler,et al.  MobileNetV2: Inverted Residuals and Linear Bottlenecks , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[4]  Bo Chen,et al.  MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications , 2017, ArXiv.