LONG delay when selecting a found object

---

@GrandyCo wrote:

Anything else I can do to fix this?

---

I just thought of a possible way. Can you test this:

First, in the options bar for the Object Selection tool, deselect Object Finder so that Photoshop doesn’t need to churn through that.

Now, use the Object Selection tool to drag a rectangle around the object you want. Is Photoshop able to isolate it faster when you give it a smaller area to analyze?

If that helps, you might also experiment with the Object Selection Tool in lasso mode (choose Lasso from the Mode menu on the options bar). Although I think Rectangle mode might be easier, because it takes a lot more time to drag a careful lasso selection around an object.

In short, try dragging the Object Selection tool around an object, instead of just hovering it.

This got me curious, because I have also seen long delays with the Object Selection tool/Object Finder at times. My personal impression is that this is one of those features that only showed up in the newest versions of Photoshop because if they had tried to add it more than a few years ago, the best hardware available then would not have run it well. In other words…

---

@GrandyCo wrote:

Is this normal?

---

From what I can tell, it probably is normal.

My brief tests below, after a cold start of Photoshop version 23.4:

Computer 1 (MacBook Pro notebook)

Intel Core i5, 4 CPU cores with (weak) Intel integrated graphics, released in 2018 (5 years old)

Object Finder time to recognize: 24 seconds

Computer 2 (MacBook Pro notebook)

Apple M1 Pro, 8 CPU cores and 10 GPU cores, released in 2021 (<1 year old)

Object Finder time to recognize: 8 seconds

My guess is that the latest (12th generation) Intel Core CPUs would perform similarly or better; the top of the line Intel processors for notebooks now have 14 CPU cores.

So it looks reasonable to assume that the time that Object Finder takes gets a lot closer to zero if the computer is very recent. It is not just that they are newer; they have more. Both Apple Silicon and Intel now include many more CPU and GPU cores than just a few years ago. And, to further accelerate machine learning features, some current processors now include hardware acceleration for machine learning, such as the Apple Neural Engine, or Deep Learning Boost in the latest Intel processors.

Given that the 4-core Intel Core i7 860 listed for your system specs was released 13 years ago (2009) if my web search was correct, it seems plausible that Object Finder could take over 2 minutes, based on the 24 seconds I got with my 5-year-old 8th gen Core i5. The Intel Core i7 860 was probably a great performer at the time. But compared to the amount of power available today, it does fall short in terms of CPU cores, GPU processing power, and won’t have any kind of hardware acceleration for machine learning.

I wouldn’t mind being wrong about this, but it looks like the performance cure for AI-type features is using hardware that’s quite recent (released after the sharp rise in cores and the addition of machine learning acceleration). That might apply to any advanced machine learning feature you try, such as the scaling in Topaz Gigapixel AI. Do you find similarly long times using the Neural Filters in Photoshop, which also use Sensei machine learning?

While your computer meets the minimum system requirements, it does not mean it will perform as well as a newer system. I would expect lags with doing something as memory intestive as the sensei selection features depending on the canvas/image size of the project. There comes a point where updating PS is going to negatively affect your work with a 13 year old processor.

What do you have for a graphics card?