Bosch AI Camera Trainer Released And Tested

By: IPVM Team, Published on Apr 09, 2019

Bosch is releasing a highly unusual new AI feature - 'Camera Trainer'. Now, coming as a standard feature in Bosch IVA/EVA analytics, one can train the camera to detect custom objects one wants.

image

While detecting a face or a person or a car is increasingly common, Bosch is allowing users to define what objects are most interesting to them - whether it is a drill or a door opening or a ladder or a buggy or whatever might be important to one's camera scene.

On the positive side, this opens up many possibilities. On the negative side, it raises the question of how well this will work and how hard it is to set it up.

IPVM tested the beta version of Camera Trainer, configuring it in multiple scenes, to see how it performed.

***** ** ********* * highly ******* *** ** feature - '****** *******'. Now, ****** ** * standard ******* ******* ***/*** *********, *** *** ***** the ****** ** ****** custom ******* *** *****.

image

***** ********* * **** or * ****** ** a *** ** ************ common, ***** ** ******** users ** ****** **** objects *** **** *********** to **** - ******* it ** * ***** or * **** ******* or * ****** ** a ***** ** ******** might ** ********* ** one's ****** *****.

** *** ******** ****, this ***** ** **** possibilities. ** *** ******** side, ** ****** *** question ** *** **** this **** **** *** how **** ** ** to *** ** **.

**** ****** *** **** version ** ****** *******, configuring ** ** ******** scenes, ** *** *** it *********.

[***************]

*******

** *** *****, ***** Camera ******* ********* ********* as *********, ******** ** to ***** ** ******** objects ***** ***** *** be ************* ** ******** analytics (**** ** ****** being ******* ** ******* left ** * ****) as **** ** ******** on *** ***** ** some *******, **** ** doors ***** ****** ** closed.

*******, ***** *** ***** key ***********:

  • ***** ********* ** ***** subjects:****** ******** *** ***** would ***** ***** *** camera ** ****** ** object **** *** *** exist, **** ** * fold ** * ******'* clothes ********** ** * drill ** * ***** appearing ** ** **** door. ***** ****** ***** not ** ******** ** adding **** ******** *******.
  • ****/*********** **** ****** *******:****** ******* ******** **** object ***********, ****** *****, and **** ****** ******* to ******** ****. **** means **** ******* ******** depth ** *** ***** (moving ****** ** *******) or ******** ********* ****** in ****** ********* ** duplicate ********* ** **** cases. *******/****** ** *********** angled ******* **** ** trained **********.
  • ******* *** ***** ***********:** *** *****, ***** 1 *** *** *****, Camera ******* ****** ** detect ******* ** ***** otherwise ******** ***** **. This ***** ** ******** with ********** ***** (******* or **) ** *** scene, *** **** ** the ********* ****** ********* include ********** **.

****: **** ** * test ** * ********** version ** ****** *******. We **** ** ****** and ****** **** ****** if *********** *********** *** found **** ** **** Bosch's ******** **** ****** Flexidome, ******** ** *****/*** 2019.

************

****** ******* ** ******** to ** ********* ** the *** ** ***** 2019. ** ** ********* on ****, *, *** 7.3 ******, ***** ********:

  • ****** ** ********* ****/****
  • ********* ** ********* ****/**** IC
  • ********* ** ********* *** ultra ****/****
  • *** ** ********* *****
  • *** ** ****** ***** (color ******* ****, *** thermal)

App/Training ********

****** ******* ******** * few ****** **** *** user, ********** ***** *****'* Configuration ******* *******:

  • ******** *******: ***** ******* are ******** ** ******* a *** ****** * specific ****** ** ** trained, **** ** * drill, * ****** ****, a ******, ***. ***** is ** ******** ****** of ******* ******** *** training.
  • ******** *******: ***** ******* are ******** ****** ********, but ******* * *** around * ****** ***** the ****** ************, ** ** ******** on * ******** ***** positive *** ******* ** as ********.
  • ****** ** ********:  *** area ***** ********* ***** place. ** *** *****, ROI *** *** ******** or ******** ******** *** was ****** **** ** narrow ********* ****.

** ****** **** ******* in *** ***** *****:

HOG **. *** **************

***** ****** **** **** initial ******* ** ****** Trainer ** ***** *** (********* ** ******** *********) *** ****** ***********. However ** *** **** key ***********, ** ** cannot ********* *****, ******* drastically ** ***** ******** (e.g. *** *****) ****** where ***** ** ******* become **** *****, *** must ** ******* ** specific ****** ******, ******* of *******.

***** *** ****-**** ***** to ********* *********** ****** Networks (****) *** ********* in *** ******, ***** should ******* **** **** of ***** ******. ** the ***-****, **** **** plan ** ********* ***-******* models *** ****** ****** classes ** ****** ********.

*** **** ******* ** the *********** ******* ***** approaches, *** ********* ********* ************.

Camera ******* ***********

*****'* ************* ******** ******* regarding ****** ******* ***********, most ************* ****/***********:

****** ******* ********* *** limited ** *** ****. Due ** ***********, ******* appear **** ******* ** the ********** **** **** the ****** *** ***** much **** *****. **** difference ** **** *** pixel ******** ****** *** image ****** ** ******* with ****** *******. **** a ****** ******** ** restricted ********** *** ********* distance, ****** ******* ********* can ** ******* ** compensate *** ***-***** **********. Objects ** *** *** distance ****** ** ******** at ***, ** **** will *** ***** ****** pixel *** *** ****** Trainer ********.

*******, ******* ***** **************, we ***** ****** ******* performed ****** **** ******** with ****** ** ******** changes ** *********** *** size, **** **** *********** than *** ************ ********** Bosch **** ***. ** discuss ***** ****** *****.

Performance ********

**** ********* *****'* ************* guidelines, *** ****** ******* was ****** ******* ** recognize **** ** ***** and **** *** **. In *** ******* *****, the ****** *** ******* to ********* **** ******* cabinets **** ******, **** an ***** ** **** open *** ****. **** that *** ****** ** able ** *********** ******* the **** ******* *** person, ****** *** ******* is ****** ******** ****** him.

Cabinet-Open_Close-Status-Detected-Properly

** ******* **** ***** configuration, *** ****** *** trained **** ****** ** all ******** **** *** closed, ** **** ** negative ******* ** * human ** ***** ** open *** ****** *****.

Training Data (Positive And Negative)

***** *****'* ******* ***** configuration, *** ****** ******* was **** ******* ** count *** **** ******** were **** ****.

Camera Trainer Data Used For Counting

** ******* *******, *** camera *** ******* ** recognize ****** ** * pegboard ****, **** ** alarm ******* **** **** were *******.

*******, **** ***** *** have **** ******, **** other ******* ************ **** as ******, ********** ****** entering *** *****, *** also ***** ***** ** the ****. ** ****** of ********* ***** *** to *** ******* ******* these ***** ******.

Humans Trigger False Detection

**** **** ***** ****** cleared ***** *** ****** left *** *****, ** if *** ****** *** to ****** * ******* number ** *******, *** rule ***** ***** ********.

Perspective *** ****** **** ******

******** ******* **** ** roughly *** **** **** and *********** ** ***** for ****** ******* ** detect ****. *** ***** object ******** *** ******** is *** **** ****** must **, ** *** Trainer's ********* *** **** is ******. **** ***** that **** ****** ** detected **** **** *** side *** ****-**, *** example, ** **** ***** change *** ******'* ****** ratio **********.

******* ** ***** ******, objects **** **** ******** shapes *** ***** **** not ******* ****. *** example, ***** *** ****** could ** ****** ******* to ****** ******* ****** in *** ***** ** view (*****), ****** **** nearer ** ******* **** the ****** ** ******** their *********** ******** ** missed **********.

Perspective Angle Impacts Accuracy

**** ******* ** ****** a ******* ******, *** camera ******** ** ** two ******** ******* **** close ** *** ******, or *** ** ***, since ** *** **** larger **** *** ***** the ****** *** ******* on.

Duplicate Objects Detected At Close Distance

**** **** ***** ********** against ****** ********* **** different ************ *** ********* to * ****** ****** Trainer ****, ** ** may ****** *********** **** looking *** ******** *******. However, ** ********** **** may ** ******* *** trained **** ******** **** an ********** ***** ** distance.

Reduced *********** ** *** *****

****** ******* *********** ** low ***** ********** (~*.***) was *******, ***** ******* objects ** *** ******* to ***** ** **** it ***** ******** ***** on ** **** *****. This ***** ** ******** using ******** ************, *** no ********* ****** ******* integrated **.

Low-Light-Drastically-Reduces-Performance-(~0 5-Lux)

VMS ***********

***** ****** ******* ** integrated ****** ***** ******** analytics, **** ******** ***** used ** ******* ** standard **** ***** (****** detected, *********, ****** *******, etc.). ******* ** ****, Camera ******* ** *********** compatible **** *** *** that ************* *********, ********* *****, *******, and *********.

*******, **** **** ***** may *** ** *********. For *******, *** "*********" rule ** ********* ** "People ********" ** ******* Security ******, ****** ** does ********* ******* *** number ** ******* *******:

VMS Integration Same As IVA

Versions ****

*** ********* ******** **** used ****** *******.

  • ***** ************* *******:*.**.****.* - **** **** this ******* ** * pre-release ****, *** ********* for ****** ***. *** non-beta ******* ** ******** in *** **** *** weeks.
  • ***** ** ********* **** VR:*.**.****

Comments (11)

*********** *******. **** ****.

*** ************* ** ****** and ********** ****** ***** thereon ** ******** ********* that **** ** ******* as * ******* **** vendors (****** "*****") ******* in ************ **** *** customer.  ***** **** **** more ******* ********* ** it ** ********* ****** and *** ******* * superior *****.  ****** **** pay ******** ******* ********* to *** *** ****** models *** **** *** inherent ****** ** **** the **** ***********, ** will ** **** ********* for ***** ** ***** as * ******* *** customization ** ****** ***** where *****'* ********* ** the ***** ******* ** training *** **** ******** the ***** ***** ** the ********'* ******.  **** super ******* ** ******** running ** *** $***,**** (*** ******'****-* ** $***,*** *** DGX-2 ** $***,***) *** the **** *** ***** it ***** ** ****** good ******, ** ***** counterproductive ** ****** ********* to ********* **** ****** when ***** *** *** tools ** ********** *** job.  * ***** ** will ** **** ********* for ***** ** **** configuration ** * *******.

**** ** **** ** can ****** * ******* left ***** ** *** checkin **** ** ** airport.

*** ****** **** ****** analytic *** **** ** place **** ***** *** a *****. *** ****** trainer ******** ** ******** to ******* * **** customized ******** *** * customer. *** ******** ****** gives *** ******* ******** and ******** ******** ** what ** ** ***** trained ** **** ***. these ******** **** ******** from *** ***** *** the ******. *** **** examples ***** *** **** accurate *** ******** **. I **** ******* ** to ********* * ***** banner *** ****** ** min ** *** *********** the ****** *** ******** other ****** ** *** scene.

***** *** ****-**** ***** to ********* *********** ****** Networks (****) *** ********* in *** ******, ***** should ******* **** **** of ***** ******.

** ***** ********** **** for *********, **** **** effectively ******* **** ********* from ******** ** ********** since ******** **** ******** thousands ** ******* ****** images. *** ******* *** implementation ** ******** ** incredibly ****** ** ***** by * ********. *** it ** ******* **** as ******** ** ****.

** ***** ********* *** have ** ****** ** the ****** ** ** the ************* ******* ******** for *** ****** ******* tab ** **** ** in ************* *******?

* **** ********* ************* Manager *.** *** **** firmware *.** ***** *** same ****** **** ** one ** *** ****** above *** * ***'* get *** ***.

**** ** * **** to *** **** ***** on *** ***** *** Site:*********** ***** ********* *.**** ******** *** **** numbers *** *** ******** are ** *** **** sheet, *** ***** ** no ****** *** *** licenses.

**#*, *** *** ** camera ******* ******** *********** ** * *******, **** *** ******* is ********* *** ****** trainer *** **** ** available ** ************* *******.

*** ****** ******* ******* does *** **** * cost ********** **** **. It ** * **** license.

*** ********** **** ******** a **** ****** ******** about ** ** ** an **** **** ** have *** **** ******* released ** ****.

***** **** **** ** ensure *** ********** *** the ***** ******* ********* needed ** *** ** the ****** *******.

**** ** *** *** time * **** ***** of *** *** *** CNN ** - ** this * ******** *** all ****?
*** **** *** * guess *** *** ** the ****** ********? *** sure * ********** *** differences.
*** * ** ****** the ** ** ***.

** **** * **** primer ****** ****:**** **** ******** ********

*** *** **** *** both ******** ****** *** video *********. *** ** an ***** **********, ********* computer ****** *********, ******* a *** (************* ****** network) ** * **** learning ****** ***** ** what **** ****** ** video ************ ******** ******* are ***** **.

**** ******** ******* *** trained **** ******** *** learn **** ******* *** based ** **** ********. HOG/computer ****** ******* *** preprogrammed **** ********** ** what *****/****** * (******/***/****/***) looks ****.

** **** ****** ********* CNNs/hardware *********, *** **** had * **** ******** Class ******* ** ****:

Read this IPVM report for free.

This article is part of IPVM's 6,367 reports, 855 tests and is only available to members. To get a one-time preview of our work, enter your work email to access the full article.

Already a member? Login here | Join now

Related Reports

Hidden Camera Detectors Tested on Nov 18, 2019
Hidden cameras are a growing problem as cameras become smaller, cheaper and easier to access. However, some companies claim to be able to detect...
Axis Tailgate Detection Tested on Jan 04, 2019
Axis is aiming to tackle tailgating, one of access control's biggest issues, with the Tailgating Detector ACAP application. This camera app claims...
Camera Height / Blind Spot Added to IPVM Camera Calculator on Oct 18, 2018
IPVM has added camera height and blind spot estimation to the Camera Calculator. This is especially helpful for those who need to mount cameras up...
Calculator - NEW Advanced Searching, Camera Favorites, Model Adding on Dec 05, 2017
The IPVM Calculator has significantly improved your ability to work with 5,500+ camera models: Advanced searching - now pick cameras based on...
Neurala Deep Learning Platform Profile on Aug 31, 2017
Find lost children? If Neurala has its way with its recent partnership with Motorola, that is the aim as shown in their video below: Neurala has...
Axis Camera Optimization App (UI Me) Tested on Aug 28, 2017
A new Axis camera app (ACAP) called UI Me has been released, offering a number of interesting features including optimization wizards and...
Avigilon Unusual Motion Detection (UMD) And Advanced Appearance Search Examined on Apr 06, 2017
Deep learning is a big trend at ISC West and Avigilon is one of the bigger names capitalizing on it. Avigilon is showing 2 upcoming...
Yi Smart Dashcam Tested on Mar 31, 2017
Mega Chinese manufacturer Yi Technology is following up on its Yi 2 'Intelligent Camera' with a dash cam. This 'Smart Dash Camera', for ~$50,...
Axis CTO: Deep Learning 'Exciting Space' But Urges Caution on Dec 27, 2016
Deep learning interest is rising in video surveillance. 2016 saw: Hikvision's partnership with Nvidia Movidius launch a neural networks chip...

Most Recent Industry Reports

Hikvision Illicitly Uses Back To The Future In Marketing on Jul 03, 2020
NBCUniversal told IPVM that Hikvision UK's ongoing coronavirus marketing campaign using NBCUniversal's assets was not allowed. Hikvision mass...
Verkada: "IPVM Should Never Be Your Source of News" on Jul 02, 2020
Verkada was unhappy with IPVM's recent coverage declaring that reading IPVM is 'not a good look' and that 'IPVM should never be your source of...
Vintra Presents FulcrumAI Face Recognition on Jul 02, 2020
Vintra presented its FulcrumAI face recognition and mask detection offering at the May 2020 IPVM Startups show. Inside this report: A...
Uniview Wrist Temperature Reader Tested on Jul 02, 2020
Uniview is promoting measuring wrist temperatures whereas most others are just offering forehead or inner canthus measurements. But how well does...
Dahua USA Admits Thermal Solutions "Qualify As Medical Devices" on Jul 02, 2020
Dahua USA has issued a press release admitting a controversial point in the industry but an obvious one to the US FDA, that the thermal temperature...
Access Control Online Show - July 2020 - With 40+ Manufacturers - Register Now on Jul 01, 2020
IPVM is excited to announce our July 2020 Access Control Show. With 40+ companies presenting across 4 days, this is a unique opportunity to hear...
Hanwha Face Mask Detection Tested on Jul 01, 2020
Face mask detection or, more specifically lack-of-face-mask detection, is an expanding offering in the midst of coronavirus. Hanwha in partnership...
UK Government Says Fever Cameras "Unsuitable" on Jul 01, 2020
The UK government's medical device regulator, MHRA, told IPVM that fever-seeking thermal cameras are "unsuitable for this purpose" and recommends...
Camera Course Summer 2020 on Jun 30, 2020
This is the only independent surveillance camera course, based on in-depth product and technology testing. Lots of manufacturer training...
Worst Over But Integrators Still Dealing With Coronavirus Problems (June Statistics) on Jun 30, 2020
While numbers of integrators very impacted by Coronavirus continue to drop, most are still moderately dealing with the pandemic's problems, June...