Transfer Learning on Convolutional Neural Networks for Dog Identification

Tu, Xinyuan; Lai, Kenneth; Yanushkevich, Svetlana

doi:10.1109/icsess.2018.8663718

Cited by 14 publications

(5 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, the Flickr-dog dataset trained only at the end of the model using a 10 k-fold cross-validation. In [34], as a r the model achieved an accuracy of 83.94% on the Flickr-dog dataset. In [31], a dog dataset was built by collecting and preprocessing a dog face dataset from the Internet a ResNet-like model suitable for the size of the collected dataset was built.…”

Section: Methodsmentioning

confidence: 93%

“…As a result, the authors achieved an accuracy of 67.6% on the Flickr-dog dataset. In [34], a dog face detection model was built by cropping dog faces from the Columbia dogs dataset and learning Faster RCNN [35]. Moreover, then, the dog face detection model cropped the dog faces in the Columbia dogs dataset and Stanford dogs dataset.…”

Section: Deep Learning On Animal Biometricsmentioning

confidence: 99%

See 1 more Smart Citation

A Methodology for Utilizing Vector Space to Improve the Performance of a Dog Face Identification Model

Yoon

Rhee

2021

Applied Sciences

View full text Add to dashboard Cite

Recent improvements in the performance of the human face recognition model have led to the development of relevant products and services. However, research in the similar field of animal face identification has remained relatively limited due to the greater diversity and complexity in shape and the lack of relevant data for animal faces such as dogs. In the face identification model using triplet loss, the length of the embedding vector is normalized by adding an L2-normalization (L2-norm) layer for using cosine-similarity-based learning. As a result, object identification depends only on the angle, and the distribution of the embedding vector is limited to the surface of a sphere with a radius of 1. This study proposes training the model from which the L2-norm layer is removed by using the triplet loss to utilize a wide vector space beyond the surface of a sphere with a radius of 1, for which a novel loss function and its two-stage learning method. The proposed method classifies the embedding vector within a space rather than on the surface, and the model’s performance is also increased. The accuracy, one-shot identification performance, and distribution of the embedding vectors are compared between the existing learning method and the proposed learning method for verification. The verification was conducted using an open-set. The resulting accuracy of 97.33% for the proposed learning method is approximately 4% greater than that of the existing learning method.

show abstract

Section: Methodsmentioning

confidence: 93%

Section: Deep Learning On Animal Biometricsmentioning

confidence: 99%

A Methodology for Utilizing Vector Space to Improve the Performance of a Dog Face Identification Model

Yoon

Rhee

2021

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…The system proposed in our paper not only identifies faces but also identifies continuous behaviors of other featured parts including information limb movement can convey. Related literature [18] used transfer learning to design a method for dog identification from shallow to deep. In this paper, different features of dogs such as eyes, noses and ears are identified.…”

Section: Literature Reviewmentioning

confidence: 99%

Multiple Feature Dependency Detection for Deep Learning Technology—Smart Pet Surveillance System Implementation

et al. 2020

View full text Add to dashboard Cite

Image identification, machine learning and deep learning technologies have been applied in various fields. However, the application of image identification currently focuses on object detection and identification in order to determine a single momentary picture. This paper not only proposes multiple feature dependency detection to identify key parts of pets (mouth and tail) but also combines the meaning of the pet’s bark (growl and cry) to identify the pet’s mood and state. Therefore, it is necessary to consider changes of pet hair and ages. To this end, we add an automatic optimization identification module subsystem to respond to changes of pet hair and ages in real time. After successfully identifying images of featured parts each time, our system captures images of the identified featured parts and stores them as effective samples for subsequent training and improving the identification ability of the system. When the identification result is transmitted to the owner each time, the owner can get the current mood and state of the pet in real time. According to the experimental results, our system can use a faster R-CNN model to improve 27.47%, 68.17% and 26.23% accuracy of traditional image identification in the mood of happy, angry and sad respectively.

show abstract

“…Indeed, Transfer Learning has found application in various works across diverse domains. In the realm of wildlife identification, this method has been utilized for fish identification in tropical waters [ 13 ], distinguishing between different dog breeds [ 14 ] and accurately identifying various bird species [ 15 ].…”

Section: Introductionmentioning

confidence: 99%

Reptile Identification for Endemic and Invasive Alien Species Using Transfer Learning Approaches

Hernández-López,

Travieso-González

2024

Sensors

View full text Add to dashboard Cite

The Canary Islands are considered a hotspot of biodiversity and have high levels of endemicity, including endemic reptile species. Nowadays, some invasive alien species of reptiles are proliferating with no control in different parts of the territory, creating a dangerous situation for the ecosystems of this archipelago. Despite the fact that the regional authorities have initiated actions to try to control the proliferation of invasive species, the problem has not been solved as it depends on sporadic sightings, and it is impossible to determine when these species appear. Since no studies for automatically identifying certain species of reptiles endemic to the Canary Islands have been found in the current state-of-the-art, from the Signals and Communications Department of the Las Palmas de Gran Canaria University (ULPGC), we consider the possibility of developing a detection system based on automatic species recognition using deep learning (DL) techniques. So this research conducts an initial identification study of some species of interest by implementing different neural network models based on transfer learning approaches. This study concludes with a comparison in which the best performance is achieved by integrating the EfficientNetV2B3 base model, which has a mean Accuracy of 98.75%.

show abstract

Transfer Learning on Convolutional Neural Networks for Dog Identification

Cited by 14 publications

References 2 publications

A Methodology for Utilizing Vector Space to Improve the Performance of a Dog Face Identification Model

A Methodology for Utilizing Vector Space to Improve the Performance of a Dog Face Identification Model

Multiple Feature Dependency Detection for Deep Learning Technology—Smart Pet Surveillance System Implementation

Reptile Identification for Endemic and Invasive Alien Species Using Transfer Learning Approaches

Contact Info

Product

Resources

About