How can I improve the performance of the ACFObjectDetector?

I'm teaching the ACFObjectDetector to detect certain indicators from an automotive instrumentcluster. While teaching the system the right indicators with the imageLabeler, the acfDetector training doesn't respond. I tried to teach it with video material and with pictures from different angles and such, but it still doesn't teach itself weak learners. The following happens:

acfDetector=trainACFObjectDetector(trainingData,'NegativeSamplesFactor',2) ACF Object Detector Training The training will take 4 stages. The model size is 53x57. Sample positive examples(~100% Completed) Compute approximation coefficients...Completed. Compute aggregated channel features...Completed. -------------------------------------------- Stage 1: Sample negative examples(~100% Completed) Compute aggregated channel features...Completed. Train classifier with 15 positive examples and 30 negative examples...Completed. The trained classifier has 19 weak learners. -------------------------------------------- Stage 2: Sample negative examples(~100% Completed) Found 30 new negative examples for training. Compute aggregated channel features...Completed. Train classifier with 15 positive examples and 30 negative examples...Completed. The trained classifier has 19 weak learners. -------------------------------------------- Stage 3: Sample negative examples(~100% Completed) Found 30 new negative examples for training. Compute aggregated channel features...Completed. Train classifier with 15 positive examples and 30 negative examples...Completed. The trained classifier has 19 weak learners. -------------------------------------------- Stage 4: Sample negative examples(~100% Completed) Found 30 new negative examples for training. Compute aggregated channel features...Completed. Train classifier with 15 positive examples and 30 negative examples...Completed. The trained classifier has 19 weak learners. -------------------------------------------- ACF object detector training is completed. Elapsed time is 102.9022 seconds.

(while it should do something like in this example: https://nl.mathworks.com/help/vision/ref/objectdetectortrainingdata.html)

Changing the samplefactor doesn't help either. Does anyone know how I can het the weak learners up? So the program doesn't trigger on the background or other indicators.

Thanks.