A cross-domain fruit classification method based on lightweight attention networks and unsupervised domain adaptation

Abstract

Image-based fruit classification offers many useful applications in industrial production and daily life, such as self-checkout in the supermarket, automatic fruit sorting and dietary guidance. However, fruit classification task will have different data distributions due to different application scenarios. One feasible solution to solve this problem is to use domain adaptation that adapts knowledge from the original training data (source domain) to the new testing data (target domain). In this paper, we propose a novel deep learning-based unsupervised domain adaptation method for cross-domain fruit classification. A hybrid attention module is proposed and added to MobileNet V3 to construct the HAM-MobileNet that can suppress the impact of complex backgrounds and extract more discriminative features. A hybrid loss function combining subdomain alignment and implicit distribution metrics is used to reduce domain discrepancy during model training and improve model classification performance. Two fruit classification datasets covering several domains are established to simulate common industrial and daily life application scenarios. We validate the proposed method on our constructed grape classification dataset and general fruit classification dataset. The experimental results show that the proposed method achieves an average accuracy of 95.0% and 93.2% on the two datasets, respectively. The classification model after domain adaptation can well overcome the domain discrepancy brought by different fruit classification scenarios. Meanwhile, the proposed datasets and method can serve as a benchmark for future cross-domain fruit classification research.

Emulating Nociceptive Receptor and LIF Neuron Behavior via ZrOx‐based Threshold Switching Memristor

“…The variation of the device was also studied, which will cause a large deviation in the artificial neural networks. [30] Both the cycle-to-cycle and device-to-device variations were tested in Figure 2b,c. In Figure 2b, the threshold switching voltages were counted from the 50 I-V cycles in Figure 2a.…”

An instance-based deep transfer learning method for quality identification of Longjing tea from multiple geographical origins

“…Hence, combining deep learning and transfer learning can not only take advantage of deep neural networks to extract discriminative semantic features but also reduce data hunger through knowledge transfer. At present, some scholars have applied basic transfer learning strategies, such as fine-tuning pre-trained models or feature extraction, to plant disease and pest detection [11,17,18], fruit classification [13] [19], and sheep facial expression classification [20]. The results show that transfer learning is an effective strategy for building high-performance classification models.…”