Search Results

Now showing 1 - 4 of 4

Some of the metrics are blocked by your
consent settings
CowXNet: An automated cow estrus detection system
(2023-01-01)
Lodkaew, Thanawat
;
Pasupa, Kitsuchart
;
Loo, Chu Kiong
Estrus detection is essential for dairy farms to take cows for artificial insemination promptly. Conventional approaches for detecting estrus cows use electronic devices attached to cows to gather data for software analysis. However, electronic devices can be costly and make a cow moody and uncomfortable while moving. In a common approach, observers detect estrus cows by observing their behaviors. However, continuous observation can easily lead to errors due to the observer's fatigue. Therefore, we designed CowXNet, an automatic estrus detection system for cows, to assist farmers to detect estrus cows. CowXNet requires only a camera attached in a pen and a computer to analyze recorded videos. CowXNet analyzes the estrus behaviors of each cow in a pen and helps farmers to identify estrus cows. To develop and evaluate CowXNet efficiently and effectively, we collected data from Chokchai Farm, the biggest dairy farm in Asia (14.65483<sup>∘</sup>N, 101.34853<sup>∘</sup>E). CowXNet has four modules: (i) cow detection uses YOLOv4 to detect cows in recorded videos; (ii) body part detection uses a convolutional neural network to estimate locations of body parts of detected cows; (iii) estrus behavior detection uses body part coordinates to extract a set of discriminative features, and a classification algorithm to detect estrus behaviors, and (iv) behavior analysis module displays estrus behavior for analysis purposes. We evaluated CowXNet for two instances: module-independent evaluation and end-to-end framework evaluation. Overall, CowXNet was promising; it correctly detected estrus behavior interval of cows 83% of cases.
Some of the metrics are blocked by your
consent settings
Hybrid Loss for Improving Classification Performance with Unbalanced Data
(2020-01-01)
Lodkaew, Thanawat
;
Pasupa, Kitsuchart
Unbalanced data is widespread in practice and presents challenges which have been widely studied in classical machine learning. A classification algorithm trained with unbalanced data is likely to be biased towards the majority class and thus show inferior performance on the minority class. To improve the performance of deep neural network (DNN) models on poorly balanced data, we hybridized two well-performing loss functions, specially designed for learning imbalanced data, mean false error and focal loss. Since mean false error can effectively balance between majority and minority classes and focal loss can reduce the contribution of unnecessary samples, which are usually samples from the majority class, which may cause a DNN model to be biased towards the majority class when learning. We show that hybridizing the two losses can improve the classification performance of the model. Our hybrid loss function was tested with unbalanced data sets, extracted from CIFAR-100 and IMDB review datasets, and showed that, overall, it performed better than mean false error or focal loss.
Some of the metrics are blocked by your
consent settings
A New Approach to Automatic Heat Detection of Cattle in Video
(2019-01-01)
Pasupa, Kitsuchart
;
Lodkaew, Thanawat
Heat detection of cattle in video is essential for dairy farm. A cow should be inseminated within a certain period of time in order for it to breed successfully. After it has given birth to a calf, it produces milk. This paper proposes the use of a set of discriminative features to detect cattle in heat, where the features were extracted from the behaviours of oestrus cow by a key-point analysis of locations of their body parts in a video. We evaluated our proposed features, in terms of the algorithm’s classification accuracy of identifying cow in heat, with several machine learning algorithms for two instances–using a global model and a number of cattle-specific models to execute the identification. It was found that Support Vector Machine with Radial Basis Function yielded a maximum accuracy of 90.0% for the global model and 92.0% for the cattle-specific models. These initial findings demonstrate that individual cows may have different oestrus behaviours, a fact that would benefit any dairy farmers. Our future development will be on a practical video monitoring and detection system of cows in heat in a dairy farm.
Some of the metrics are blocked by your
consent settings
Fashion finder: A system for locating online stores on instagram from product images
(2018-11-13)
Lodkaew, Thanawat
;
Supsohmboon, Weeruhputt
;
Pasupa, Kitsuchart
;
Loo, Chu Kiong
Searching for a fashion or clothing shops that carry either the piece that a consumer desires or a similar one can be troublesome due to lack of complete information, e.g., lack of the name of the brand or the location of a shop. A consumer may also waste a lot of time crawling in an online store. Many online stores utilises a social network called 'Instagram' as a digital advertising platforms. In most of online stores, a user can simply use a keyword to search for items. Unfortunately, this cannot be done in Instagram. A consumer may search for a store on Instagram by using Instagram's ID assuming that the store is known. However, he or she needs to search each post one by one until he or she finds the desired product which may take a long time. A way to directly access these stores is needed. Therefore, we propose a platform called Fashion Finder that can assist consumers in their search for Instagram's online stores that sell their desired piece or an equivalent one. Fashion Finder uses a deep learning algorithm to be able to do so. It is simple to use, fast and a shop owner can add more items to the platform by themselves. The experiment shows that our proposed framework outperforms the conventional approach on the Colorful Fashion Parsing dataset.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results