3 Mistakes you should avoid | Computer Vision

Today we will see the 3 most common mistakes you should avoid when working on your computer vision project.

How do I know what to recommend?
Because since I opened my Youtube channel and developed projects for many clients, I have received several comments and feedback. Based on these I have identified some patterns and thanks to this I can indicate which are 3 main errors.

If you find yourself in the descriptions, work on it and you will get great results.

1. Skip the basics

The main mistake is that people skip learning the basics. Why does this happen?

You’ve lack of time

People don’t have time to study the basics or all the libraries. It is thought to proceed directly with the project trying to get by.

You’re in a hurry

They have a project to be delivered in a very short time and they are in a hurry

You’re lazy

Why should they study all the material or do a lot of exercises?

The main question is why do I have to waste time, even if I don’t feel like it, to study the basics and apply myself in exercises if I’m finally on the way to finishing my project?
You are probably very lucky and your computer vision project works perfectly but what if you miss that little change or tweak to finish it and don’t know how to proceed?

But above all if you use ready-made codes you have no idea what to change, you have no idea what you did and you get a lot of frustration with the only result that in the end, you give up everything. It is much more common than one might think. You just give up and think it’s too complicated for you or it’s just not your way and many computer vision careers just end up like that.

Don’t skip the basics

What do I mean by basics of Computer Vision?

Studying the official documentation takes a long time. I studied it from the official guide and from my experience I can say that to carry out a computer vision project, in most cases, only knowing a few but very important concepts is enough to have sufficient bases to do most of the projects.

For example, in one of my courses dedicated to Object Recognition with deep learning I packed the basics in one module with 2 hours of video lessons. Because you don’t need really to know more than that to do most of the projects, still these couple of hours of content will help you tremendously and give the confidence to do further projects.

In any case, remember: The better you know the basics, the more confidence you will have to implement more advanced features.

2. Copy/Paste source code

There is nothing wrong with that, much of the source code is Open Source freely downloadable, and available for you to do whatever you want with it. Why do people copy and paste other source codes?

You think it’s too long

Solving the same problem requires too many lines of code and therefore you believe it is easier to use a ready-made one.

You think it’s too complex

Sometimes the code is too complex. Often the project you want to make requires many files, thousands of lines of code and you don’t have enough skills for it. But it may happen that to solve your problem all that code is not necessary and could be solved with a few lines and in a simpler way.

You believe you don’t have enough skills to do it yourself

It doesn’t happen often, many times in my video I explain some basic python code and it may happen that there are some people who ask very basic questions that show me that they haven’t watched the video in full.
My videos on average last between 30-40 minutes and I generally explain line by line. It happens that someone asks me things already explained in the video.
It takes me a long time to make the videos and try to explain in the best and simple way what I do in the videos, so I recommend you put some effort into it for 40 minutes so you can save yourself a lot of stress and a lot of time by understanding exactly what is going on in the code.

3. Over-Complicate

Finding a simple solution to a complex problem.

You believe there can’t be an easy solutions for complex problems

Sometimes you work hard on a project that wastes you a lot of time and in the end, you find a simple solution. So simple that we do not accept that we have lost so much useless when the solution was at hand.

You’re chasing latest researches

People are always looking for the latest algorithm that maybe allows you to improve performance by 5% but there is a big drawback. For example, the libraries are not tested enough and will give more problems after or there is no compatibility between the versions of python and libraries.

Don’t always go to the latest search, always look for the simplest possible implementation.

For example, if I try to identify a phone first I get something that works then I proceed with improving the algorithm or finding more powerful hardware to increase accuracy. All improvements in speed and accuracy must be done only after we achieved our main goal.

You’re working on unnecessary things

When you are working with your object detection project and before reaching the result you waste time improving the graphic aspects but at the end of the day you are not able to complete your project because you have put all your effort into things that do not give you a concrete result.

A list of necessary things could help you and only when you have achieved the result can you go on the graphical improvements

Blueprint

Learn to build Computer Vision Software easily and efficiently.

This is a FREE Workshop where I'm going to break down the 4 steps that are necessary to build software to detect and track any object.

Sign UP for FREE