What are the most significant risks that could jeopardize the success of the project? How are these risks being managed? What contingency plans are ready?
The most significant risk that could jeopardize the success of the project is the usage model accuracy. We were able to successfully train our clothing and color classification models, but our usage model is struggling to produce useful results, as discussed in our weekly meeting. This risk is being managed by our feature which allows users to correct and confirm the labels of each piece of classified clothing.
Were any changes made to the existing design of the system (requirements, block diagram, system spec, etc? Why was this change necessary, what costs does the change incur, and how will these costs be mitigated going forward?
No changes were made.
Provide an updated schedule if changes have occurred.
Our schedule has not changed.
This is also the place to put some photos of your progress or to brag about a component you got working.
Photos of our progress are located in our individual status reports.
Now that you have some portions of your project built, and entering into the verification and validation phase of your project, provide a comprehensive update on what tests you have run or are planning to run. In particular, how will you analyze the anticipated measured results to verify your contribution to the project meets the engineering design requirements or the use case requirements?
Validation is usually related to your overall project and is likely to be discussed in your team reports.
Our verification tests are discussed in our individual status reports.
Our validation tests will be ensured that our camera can take sufficiently clear images of our clothing for our classification models. Our models expect 224×224 images, and our camera can take images with a variety of resolutions, all sufficiently high enough to be scaled down to 224×224, so our camera will meet our requirements. Our camera also allows us to have higher resolution images to display to the user, since 224×224 images are grainy to the human eye.
Our validation tests will also verify that our classification models are compatible with the Jetson and that the Jetson is capable of predicting a class with each model in under 1 second. We have already validated that our .keras models are compatible with the Jetson by uploading them onto it. We will validate that our Jetson can run inference in under a second by using the camera to take at least 10 images of clothes from our own closet and feeding them through each of our three models. We will time how long it takes for each model to classify the image, and determine if the average inference time meets our 1 second threshold.
We will also validate that the Jetson can accurately send the classification results to the backend for storage in the database, and image folder. After the model classifies images of clothing, the Jetson will send the data as a POST request to the backend. We will validate that the backend will parse the data, add the clothing information to the database, and add the image file to the images folder. For our metrics, we will verify that the data in the database accurately matches the information sent from the Jetson. We will also confirm that the image is saved under the correct image file name, that matches the entry name for that item in the database. For speed, we will measure the time taken from sending the POST request to receiving confirmation from the backend. The goal is to meet a latency requirement of under 4 seconds for classification and storage.
Our validation tests will verify that the application accurately generates outfits based on the user’s request, and weather, along with user preferences and wardrobe data stored in the database. The outfits should meet our speed and accuracy use case requirements, and should be displayed on the frontend.
