EEE-6512EEL-4930 Image Processing and Computer Vision Spring 2024 Homework #7 (Optional)
March 21, 2024
Due: April 22, 2024, 11:59 PM
This assignment should be completed individually by the student. Proper citation should be provided for any references used.
Please read the requirements carefully. Solutions that do not follow the provided specifications will not receive credit. You are free to use any built-in/toolbox functions within MATLAB to accomplish this task, except functions from the deep learning toolbox. Data and background were taken from the Sign Language MNIST Kaggle dataset page [1].
Background: The original MNIST image dataset of handwritten digits is a popular benchmark for image-based artificial intelligence methods, but researchers have renewed efforts to update it and develop drop-in replacements that are more challenging for computer vision and original for real-world applications. American Sign Language (ASL) MNIST is one such dataset, consisting of images of hand gestures that represent a multi-class problem with 24 classes of letters (excluding J and Z, which require motion). This has applications in live ASL/spoken word translation. See provided asl_reference.png for the ASL letters. For more information, please see [1].
Data: The dataset format is patterned to match closely with the classic MNIST. Data is stored in CSV format with:
1. a label (0-25) as a one-to-one map for each alphabetic letter A (and no cases for 9=J or 25=Z because of gesture motions)
2. pixel1, pixel2, .. pixel784 which represent a single image. Data was preprocessed by cropping to hands-only, gray-scaling to uint8 bit depth, and resizing to a 28x28 pixel image.
3. For this assignment, we are only concerned with the letters A-D, inclusive. Please use the provided asl_mnist.csv.
Challenge: Write a function, myASLTranslate, which:
• accepts a single 28x28 uint8 greyscale image and returns a single character, either “A”,
“B”, “C”, or “D”.
• You must:
o Use at least one filter on the grayscale image
o Use at least one morphological image processing operation
o Use at least one region feature from section Chapter 12
o Include in your report the accuracy of your function on all data in asl_mnist.csv
• Note: Your function will be tested on 25 randomly sampled images taken from the provided asl_mnist.csv. Code must achieve at least 80% accuracy on the sampled dataset to receive credit.
To receive full credit, you should submit two files. 1.) A document containing an explanation of how your code works, (.DOC, .DOCX, or PDF file) 2.) An M-file containing commented MATLAB code for the program myASLTranslate. Students should ensure that their M-files execute without errors to avoid receiving point deductions.
References
请加QQ:99515681 邮箱:99515681@qq.com WX:codinghelp
- 赛诺威盛:大孔径专科化CT领航者
- 网易硬刚腾讯 两大游戏玩家之间的口水仗不断
- 全球“最独特”的一台华为 nova 6 5G 版手机是什么样子的?
- 拼多多抖音淘宝京东,谁是真低价?
- 老杨第一次再度抓握住一瓶水,他由此产生了新的憧憬
- 丰田章男称未来依然需要内燃机 已经启动电动机新项目
- B站更新决策机构名单:共有 29 名掌权管理者,包括陈睿、徐逸、李旎、樊欣等人
- 苹果罕见大降价,华为的压力给到了?
- 三明列东又有房子要拆迁!住这里的人要发了!
- 放大招后,广州又忍不住了…
- 私募积极加仓,百亿股票私募仓位指数创出近八周新高
- 他,传闻中马云最想见的人
- 升级的脉脉,正在以招聘业务铺开商业化版图
- 如何经营一家好企业,需要具备什么要素特点
- 智慧驱动 共创未来| 东芝硬盘创新数据存储技术