# How can a NN guess the velocity of an object from a video source?

Data Science Asked by Evan P on November 5, 2020

In the case of a video stream, I’d like to detect the speed (an approximation) of an object, that is moving. What would be the best approach to take?

Method 1: I could feed Frame images to my CNN and pinpoint the object in the frame. Then calculate the difference of the position boxes.

Frame 1: [  o     ] ---> NN --> 3
Frame 2: [      o ] ---> NN --> 7


7-3 = 4 -> so dX = +4 between 1 frame, so I can estimate the dT (if I know the sampling rate of the frames)

Method 2: Is there any way where the NN can take as a feed-input the context of the previous frame and make the calculation itself?

video lapse 1sec: [  oooo ] ---> NN --> 4m/sec


Method 3: What if I can control the shutter speed of my camera, could I calculate the velocity of an object by the motion blur?

Frame image: [   ---o ] ---> NN --> 4m/\$shutterSpeed


Any relevant resource to read would help a lot.

That is not possible without external labeled data (i.e., each video would have to be labeled with the object speed). Without labels, the neural network would be unable to learn the object's velocity because the object might be moving in any direction relative to the camera. For example, an object directly away would be getting smaller without motion blur.

Answered by Brian Spiering on November 5, 2020

## Related Questions

### Using class weights in Keras with multiple binary outputs which are not simply one-hot-encoded

1  Asked on November 16, 2020 by andr-christoffer-andersen

### How to set a class_weight Dictionary for Random Forest?

1  Asked on November 15, 2020 by fega_zero

### Forecasting Consumption for Multiple Products for Multiple Regions

1  Asked on November 15, 2020 by aatmaram

### Faster grid search with small dataset to derive best params instead of full dataset?

2  Asked on November 15, 2020

### How to combine rows after Pandas Groupby function

1  Asked on November 14, 2020

### Finding aggregated information of data

1  Asked on November 14, 2020 by pnv

### Bag-of-words model : Boolean vs. TF-IDF

1  Asked on November 12, 2020 by qualia-universe

### How can I get probabilities of next word with ELMO?

2  Asked on November 11, 2020 by natalia-bogatykh

### What is the definition of imbalanced data set

2  Asked on November 11, 2020 by alex-p

### Comparing multi-class vs. binary classifiers in predicting a single class

1  Asked on November 10, 2020 by pedram

### Determining which categorical data is beneficial in predictive modelling

1  Asked on November 10, 2020

### Applying machine learning to cache algorithm

1  Asked on November 9, 2020 by lem0n

### How do I prepare data in which each output row depends on multiple input rows?

1  Asked on November 9, 2020

### Extracting sections from document based on list of keywords – Python

1  Asked on November 8, 2020 by jemar-villareal

### Solving the dual problem of boosting using column generation

0  Asked on November 8, 2020 by code-pope

### Does using user-specific accumulative variables causes data leakage?

2  Asked on November 7, 2020 by corel

### How can you include information not present in an image for neural networks?

5  Asked on November 7, 2020

### Predictive power of a dataset

2  Asked on November 7, 2020 by oklm