全國中小學科展

電腦科學與資訊工程

以時頻分析方法量化自主神經活性

許多疾病與自主神經活性息息相關,例如睡眠呼吸中止症。現有的分析方法,例如頻率分析法,將心律變異訊號(HRV)訊號假定為平穩過程,使得我們只能得到這段訊號內的平均值,無法知道其即時資訊。本研究成功以短時距時頻分析轉換(Short-time Fourier transform)配合重新分佈方法(reassignment method)來分析自主神經系統的活性。藉由觀察隨時間變化的頻率,我們可以得知即時的交感神經與副交感神經活動。本研究可延伸應用於睡眠呼吸中止症的診斷。

運用資訊熵以及動態結構描述的兩種象棋開局庫表示法之探討與實作

象棋程式主要都是使用評估函數搜索,但是搜索有一定的不穩定性。因此,它們大多會使用開局庫,避免在開局時搜索出不好的棋步,進而影響勝負。不過,一個完整、強大的開局庫通常會佔用很大的空間,而且在檢查、維護與修改上,均有一定的難度。 本研究分別運用靜態和動態結構兩種不同的方式來表示象棋盤面,藉此減少表示一個象棋局面所需要的空間。減少它的空間後,一個相同大小的開局庫中不但可以放入更多資料,增強棋力,而且在搜尋、檢查與修改上,都能夠顯著的增加效率。 使用這兩種新的編碼方法,可以把含有超過十萬局棋(上千萬個局面)的開局庫編碼成6MB的大小。而這個開局庫,不像傳統的開局庫,只要檢查到一個不好的著法,就可以迅速的把包含這個著法的局面全部刪除。另外,從這個開局庫也可以直接取得開局的棋譜。

運用GAN實現字體風格轉換

本研究以實作字體風格轉換的生成對抗網路模型為動機,將Conditional GAN當作模型的基礎,探討pix2pix模型及其他研究的一些方法對模型會產生甚麼影響,以得出能最優化預測成效的深度學習模型。 首先進行的是前處理的步驟,將字體的truetype檔案轉換成模型輸入的jpeg檔,再以生成器(Generator)和判別器(Discriminator)建立Condional GAN的基礎模型,然後探討加入U-Net、Category Embedding等方法,以及訓練資料集大小對模型造成的影響,最後實作整合的pix2pix模型和CycleGAN模型進行比較。 經過實驗後發現,U-Net和Category Embedding都對模型的預測成果有所幫助。另外,對相似的字體而言,pix2pix的效果較好,而對兩種風格差異較大的字體則需用Category Embedding的方式,融入更多字體進行訓練以達到更好的成效。

看見拔河中的力

本設計利用了應變規(Strain Gauge)的原理去製作一台可以量測拔河時繩子兩端的施加力量值,本設計採用的原理是當應變規受力時其電阻值會因為不同的力量而產生不同的應變的變化。進一步地,本設計的準確性也進行了校正。透過拔河繩拉力所產生的應變造成電阻變化,來轉換成力量數值,而利用自已設計製造的夾具,放置在拔河繩上並固定在繩子的兩側,分析其應力應變和受力狀況來測出力量值。 本設計只用簡單的應變規、電子顯示器和電子零件,低成本、簡單構造的設計來達成本研究之目的,來解決拔河過程中比賽的兩隊平衡力之顯示,使拔河更增加樂趣,增加拔河的的趣味性、數值化。

英文句子依閱讀程度進行簡化之研究

英文句子簡化是一項單語言句子轉換的任務,其中一句複雜的句子會轉換為一句或多句的簡單句子。相較於過去研究學者著重於研究如何優化句子簡化的結果,如何將一句英文句子依閱讀程度簡化為不同簡單程度的簡化句是一項自然語言處理方面嶄新的研究領域。本研究首先訂定英文分級標準,整合歐洲(CEFR)與台灣(LTTC)母語非英語國家機構對英文的分級標準,將英文分為三種難易程度,並依此將Wekipedia及Newsela的簡化前-簡化後平行語料重新刪整為三種目標程度等級的平行語料庫。另一方面,運用已發展成熟的Seq2seq簡化模型,創造一個多解碼器模型,分別依據目標程度不同的訓練資料集訓練三種解碼器。在BLEU、SARI指標以及Coverage計算下,本研究結果相較於相關研究可展現出優異成果。

DetectTimely

This research project focuses on developing a web-based multi-platform solution for augmenting prognostic strategies to diagnose breast cancer (BC), from a variety of different tests, including histology, mammography, cytopathology, and fine-needle aspiration cytology, all in an automated fashion. The respective application utilizes tensor-based data representations and deep learning architectural algorithms, to produce optimized models for the prediction of novel instances against each of these medical tests. This system has been designed in a way that all of its computation can be integrated seamlessly into a clinical setting, without posing any disruption to a clinician’s productivity or workflow, but rather an enhancement of their capabilities. This software can make the diagnostic process automated, standardized, faster, and even more accurate than current benchmarks achieved by both pathologists, and radiologists, which makes it invaluable from a clinical standpoint to make well-informed diagnostic decisions with nominal resources.

Utilizing Computer Vision And Machine Learning Algorithms To Control Smart Systems Helping Physically Disabled People.

About 15% of the world's population lives with some form of disability, of whom 2-4% experience significant difficulties in functioning. The global disability prevalence is higher than previous WHO estimates, which date from the 1970s and suggested a figure of around 10%. This global estimate for disability is on the rise due to population ageing and the rapid spread of chronic diseases, as well as improvements in the methodologies used to measure disability. This research deals specifically with the physically disabled and often people with physical disabilities feel frustrated because they cannot do activities such as: playing sports and doing exercise. Having a physical disability also changes the way a person lives their life. They may find their life changes and activities they had previously included as part of their daily routine such as brushing their teeth, washing and doing household chores suddenly become a huge effort and many people require another person's help to carry out these activities. Also, they suffer from three basic challenges like; education, economic and, communication. Firstly, Education: The results of the investigation revealed that the physically handicapped. They face a lot of problems while studying they can't learn as the normal ones and they needs someone to help in learning. Secondly, Economic: they can't work and achieve income to help in his practical life. And finally Communication: they can't communicate with others because of his disability.

Enhancement of Online Stochastic Gradient Descent using Backward Queried Images

Stochastic gradient descent (SGD) is one of the preferred online optimization algorithms. However, one of its major drawbacks is its predisposition to forgetting previous data when optimizing through a data stream, also known as catastrophic interference. In this project, we attempt to mitigate this drawback by proposing a new low-cost approach which incorporates backward queried images with SGD during online training. Under this new approach, we propose that for every new training sample through the data stream, the neural network is optimized using the corresponding backward queried image from the initial dataset. After compiling the accuracy of the proposed method and SGD under a data-stream of 50,000 training cases with 10,000 test cases and comparing our algorithm to SGD, we see substantial improvements in the performance of the neural network with two different MNIST datasets (Fashion and Kuzushiji), classifying the MNIST datasets at a high accuracy for the mean, minimum, lower quartile, median, and upper quartile, while maintaining lower standard deviation in performance, demonstrating that our proposed algorithm can be a potential alternative to online SGD.

A Person Re-identification based Misidentification-proof Person Following Service Robot

Two years ago, I attended a robot contest, in which one of the missions required the robot to follow the pedestrian to complete the task. At that time, I used their demo program to complete the task. Not long after, I found two main issues: 1. The program follows the closest point read by the depth camera, which if I walk close to a wall next to, the robot may likely ‘follow’ the wall. 2. Not to mention if another pedestrian crosses between the robot and the target. Regarding these two issues, I decided to improve it. We’ve designed a procedure of using YOLO Object Detection and Person re-identification to re-identify the target for continuous following.

A Person Re-identification based Misidentification-proof Person Following Service Robot

Two years ago, I attended a robot contest, in which one of the missions required the robot to follow the pedestrian to complete the task. At that time, I used their demo program to complete the task. Not long after, I found two main issues: 1. The program follows the closest point read by the depth camera, which if I walk close to a wall next to, the robot may likely ‘follow’ the wall. 2. Not to mention if another pedestrian crosses between the robot and the target. Regarding these two issues, I decided to improve it. We’ve designed a procedure of using YOLO Object Detection and Person re-identification to re-identify the target for continuous following.