Anomalous events detection in real-world video scenes is a challenging problem owing to the complexity of anomaly and the untidy backgrounds and objects in the scenes. Although there are already many studies on dealing with this problem using deep neural networks, very little literature aims for real-time detection of the anomalous behavior of fish. This paper presents an underwater fish anomalous behavior detection method by combining deep learning object detection, DCG (Directed Cycle Graph), fish tracking, and DTW (Dynamic Time Warping). The method is useful for detecting the biological anomalous behavior of underwater fish in advance so that early countermeasures can be planned and executed. Also, through post-analysis it is possible to access the cause of diseases or death, so as to reduce unnecessary loss, facilitate precision breeding selection, and achieve ecological conservation education as well. A smart aquaculture system incorporating the proposed method and IoT sensors allows extensive data collection during the system operation in various farming fields, thus enabling to develop optimal culturing conditions, both are particularly useful for researchers and the aquaculture industry.

Seafood, especially the traditional one in Taiwan, is rarely sourced from a fixed species and routinely from similar species depending on their availability. Hence, the species composition of seafood can be complicated. While a DNA-based approach has been routinely utilized for species identification, a large scale of seafood identification in fish markets and restaurants could be challenging (e.g., elevated cost and time-consuming only for a limited number of species identification). In the present study, we aimed to identify the majority of fish species potentially consumed in fish markets and nearby seafood restaurants using environmental DNA (eDNA) metabarcoding. Four eDNA samplings from a local fish market and nearby seafood restaurants were conducted using Sterivex cartridges. Nineteen universal primers previously validated for fish species identification were utilized to amplify the fragments of mitochondrial DNA (12S, COI, ND5) of species in eDNA samples and sequenced with NovaSeq 6000 sequencing. A total of 153 fish species have been identified based on 417 fish related operational taxonomic units (OTUs) generated from 50,534,995 reads. Principal Coordinate Analysis (PCoA) further showed the differences in fish species between the sampling times and sampling sites. Of these fish species, 22 chondrichthyan fish, 14 Anguilliformes species, and 15 Serranidae species were respectively associated with smoked sharks, braised moray eels, and grouper fish soups. To our best knowledge, this work represents the first study to demonstrate the feasibility of a large scale of seafood identification using eDNA metabarcoding approach. Our findings also imply the species diversity in traditional seafood might be seriously underestimated and crucial for the conservation and management of marine resources.