Faulty Hikvision fever screening systems are spreading despite prominent UK experts in government and academia criticizing fever cams as 'more marketing than medicine' and recommending they not be purchased.
This 1-minute video overviews the situation:
In this post, we examine why these systems risk public health, especially with COVID-19 cases in the UK (and Europe generally) are soaring.
Hikvision UK Healthcare Deployments
IPVM identified 7 UK hospitals and nursing homes (care homes) deploying Hikvision fever screening solutions:
*** ******* ***** *** *** ** these ******** *****more ** ********* **** ******* ********. Infrared screening for temperature results in large numbers of false positives, either offering false reassurance or unnecessary alarm – and potential exclusion of the person from work or leisure activities. The nature of this testing risks public embarrassment and confidentiality when used in the mass setting. Temperature screening is not ******** *** ****** ********* *** ** ****. [emphasis added]
There ** *** ****** ******** ** ********* *** *** ** *** [infrared thermal imaging] for the prevention of COVID-19 transmission in health and care settings.
******** ********** **** ******** *****-** ******** (i.e. ** *****) *** ***** ** which ***** *** ****** *** *** still **********, **** *** ** ********** by ***** *********. ****, ******** **** the ****** ********* ***********, *** *** impact ** *********** ********** ** ***********, indicate **** ****** *** **** ********should *** ********* ******** ******* ******* as a means of detecting COVID-19 patients. [emphasis added]
Hikvision ** ********
**** ******** ********* ** *** ******** with ***** ******* *********** *** ********* did *** *******.
@Charles - I feel the term ‘faulty’ is misused here, perhaps the term ‘incorrectly deployed’ is more correct?
Clearly there is likely an issue with the way these systems are installed and used.
There also seems like a training and sys op issue on some of these too...
There could be a robust process in place but your enquiries are likely to be received and answered by the marketing/press officer rather than someone with more knowledge?
I feel the term ‘faulty’ is misused here, perhaps the term ‘incorrectly deployed’ is more correct?
These systems are faulty in multiple ways - yes, incorrectly deployed but also fundamentally faulty since Hikvision only measures forehead temperature and also faulty in missing elevated temperatures (IPVM testing).
your enquiries are likely to be received and answered by the marketing/press officer rather than someone with more knowledge?
That said, I agree to your broader point. If there is someone within Hikvision who will listen, what Hikvision is doing here is dangerous and for a company of Hikvision's size and stature, self-destructive.
Totally agreed. In my experience, the clients know exactly what they a buying when described to them. Its a small cog in a Covid strategy that neither begins or ends with an elevated temperature camera. Most of the installations in this report have indicated the exact same understanding. No-one expects a camera to "detect" Covid when correctly installed, merely a first indicator of potential risk - kind of like walking in a building without a face mask, would draw attention to a perceived risk - an elevated temperature indication would do just the same.
A good indication of a poor installation by the installer is the incorrect date format on the cameras as seen on the Carlton house installation.
The article skirts around the fact that asymptomatic carriers will pass every elevated temperature camera test - so the system ability to also identify no face mask being worn, is tangible and useful.
The article skirts around the fact that asymptomatic carriers will pass every elevated temperature camera test - so the system ability to also identify no face mask being worn, is tangible and useful.
Face mask detection is only useful if there is no one at the check point but since, by definition, these systems need to be staffed for immediate response to people with elevated temperatures, there is not much benefit for what a human can easily see.
John, while by definition these systems need to be manned, we have had a number of customers who want the system at smaller locations, tied to the access control system, with a positive result locking them out (in some cases the person starts the day locked out and if they don't get a negative at the camera they can't come in). They are not looking to have someone monitoring the system on a regular basis, only as needed.
I do agree that having face mask detection on these cameras doesn't do much, because once I pass the camera I can just take the mask off if I want to. It's kind of a boondoggle because the existing cameras on your surveillance system aren't going to be at the right angle to reliably detect face masks, so you end up putting in even more cameras to catch an outlier or two.
I would agree with my colleague, using correct language is important to credibility and the national inquirer style headlines I've seen around the subject lately undermine the valuable and well thought out reporting and writing that is actually done in the story.
To the article's point, it amazes me the number of end users out there that have spent lots of money on these "solutions" but did not do the proper research to make the decision or ensure they are properly being used. This laziness on the part of the customer is only amplified by the predatory tactics that some of those in this market appear to be using.
I'd start by saying the headline is a bit ambiguous, which is faulty the camera or the deployment? Since the article is actually about the deployment, a better choice of words would be "poorly executed". The deployments outlined by IPVM are dangerous regardless of the camera used, if it were FLIR we'd have the same issues.
The only specific reference to Hikvision itself is the sentence covering forehead screening with a link to an article discussing why this is an issue. If one clicked on the article wanting to read about how the Hik camera is faulty they'd have to click through and even that article doesn't provide the complete picture as you clarified.
I get that the word faulty helps trigger the emotions that lead to a click but a more factually accurate title with some of the same factor would be "More dangerous Hikvision fever camera deployments in the UK".
The only specific reference to Hikvision itself is the sentence covering forehead screening
We have a whole section covering Hikvision marketing this to hospitals (i.e.., "Hikvision UK Promotes Hospital & Care Home Fever Systems"). And to be clear, Charles has been researching this for 2 months and Hikvision was the system that came back again and again.
The deployments outlined by IPVM are dangerous regardless of the camera used, if it were FLIR we'd have the same issues.
No, you would not because the way FLIR is designed, the issues with this setup would immediately be found because FLIR does straight thermography. Hikvision does not. They smooth out temperatures reported that mislead users:
The smoothing out means that when Hikvision gets a bad read (e.g., someone walks past with a hat on), Hikvision regularly reports a near 'normal' temperature misleading users into thinking the system 'works' whereas FLIR regularly reports obviously low temperatures or no reading at all. Does that make sense to you?
I agree with others, the term faulty should not have been used. Criminal, or criminally negligent on the other hand, these terms do seem accurate to me.
I think it depends how they are used and sold. Reducing Covid-19 is clearly directly related to mitigation strategies implemented. No single strategy (i.e. hand washing) is a panacea in itself. It is the combination of strategies that reduces, not eliminates, risk.
An infectious disease specialist explained it to us this way - think of each mitigation strategy (hand washing, social distancing and masks) as layers of swiss cheese. Individually they all have holes in them. But as you stack them together, the holes get smaller but not completely eliminated. I see elevated skin temperature detection as another layer of swiss cheese.
Manufacturers, dealers and end users need to be cautious of how they are promoting "fever" cameras. It should be promoted as part of an overall mitigation strategy as another element of mitigation. If St. Monica's Trust, for example, is promoting it specifically "to keep residents safe" that is irresponsible. But if they are implementing it with all of the aforementioned strategies, then I don't see a problem with it.
I think this is an important conversation to have though especially to protect manufacturer's from themselves. If such reporting can force manufacturer's to change how they are promoting, and be more responsible with their marketing, maybe dealers will follow? It would be a shame if a manufacturer just saw this as an attack and didn't look beyond themselves or their bottom line to actually see the point and adjust how they promote their solutions.
Piling on the mitigation strategies is great. This is common in real world solutions. Think drivers licenses, road safety laws, vehicle crash impacts, and finally seat belts.
The difference here is that this implementation of the Hikvision cameras is that they just don't work. It's not a mitigation strategy its a false sense of security- which I'd argue is the exact opposite of the intended outcome.
I prefer the onion layers analogy myself as I use it when discussing security. When properly executed, EST screening avoids many of the issues brought up by Oxford and HPS as citied by the article.
The big argument against at least for public facilities is that another mitigation strategy is the questionnaire, administered by lobby staff. Unless the company is thinking they will want to maintain EST screening beyond Covid, it's going to be cheaper to have lobby staff record temperatures using a non-contact thermometer while they are managing the questionaire.
On a Facebook group regarding this article, they are claiming that Charles’ email gave them [not Hikvision] less than 4 hours to respond. It would disappoint me if it were true as I hope that your team have more journalistic professionalism.
I will ask Charles the exact about of time between Charles's email and our publication but this is a red herring at best. Hikvision has known about our concerns for months and has not responded. Also, if Hikvision somehow, after months of not commenting has a response, they are welcome to send it.
Why can't you Hikvision partners just face the reality that Hikvision is happy to make money off these deployments despite the obvious risks and opposition from UK health experts?
1. I’m not surprised by you attacking Hikvision (as you have also featured other manufacturers)
2. There are images posted of the email from Charles to one of the companies featured in the article. I am seeking to learn if this is true or has been altered.
3. The Facebook group is a closed group and I do not have permission to republish content.
4. I’m not particularly in allegiance with any manufacturer (as we have discussed previously) and I work for multiple companies.
5. If a system is deployed correctly and within the manufacturer guidelines then your issue is with the manufacturer. So sue them or get the feds/law enforcement to order them to stop selling / recall the products. If they are as bad as IPVM imply then surely they would be banned from sale?
If the system has not been installed in accordance with the manufacturer guidelines then the issue is with the installer.
Picking on end users is not the right thing to do. The issues with most of these systems seems to be the way they are deployed which is a training / design issue.
Correct height, angle, blackbody, non-thermally reflective backdrop, stable environment, sunlight free, correctly calibrated etc should all be done. If installers can’t follow these simple guidelines or it’s not possible to comply because the installation environment is not suitable, then these products should not be fitted!
I like the tests you do (as they are straight reporting) but take issue with the awful sensationalistic ‘Daily Star’ style headlines. By beating on Hikvision so much, all you’re doing is cementing people’s support for them as the underdog...
If they are as bad as IPVM imply then surely they would be banned from sale?
Hikvision fever cameras did not exist at the beginning of this year. It takes a while for governments and scientists to catch up to companies that rushed solutions without appropriate care or consideration. Our research is helping to inform the public and policy makers of these concerns.
If the system has not been installed in accordance with the manufacturer guidelines then the issue is with the installer.
Hikvision went out and told installers and end users that they could do fever screening with people with hats:
They told installers and end users that they could do up to 30 people at a time:
Take responsibility Hikvision. They rushed out, wanted to make money in a deadly pandemic, put out advertising about fever detection, detecting people with hats on (when they knew they only did forehead detection), detecting 30 people at a time, at front entrances, etc.
Now you all want to blame the installers? Pathetic.
Hi UI#1, thanks for your comment. I emailed Hikvision's Europe-based communications team and their UK team at 10:28am UK time and prominently wrote that we'd be publishing our article at 2pm UK time, and if they have any comment "please let me know ASAP". I received no reply at any point.
Typically, what is done when the subject of a story needs more time to respond is 1) they will say so and 2) we will grant them that time. However, Hikvision ignored our request for comment altogether.
Overall, though, I agree that for this UK article, I gave a deadline that was too short. Typically I do give substantially more time for comments. To avoid any future controversy, in the future I'll strive to give more advance notice (at least one business day).
I'd like to note for context that Hikvision almost never gives me comments for articles. However, I know they read my emails because they have quietly deleted problematic content without responding, e.g. for this story on Hikvision selling a Uyghur-detecting camera on its China website, we gave them ample time to respond but they never did, instead they quietly deleted the listing as soon as we requested comment and then just ignored us.
I was specifically talking about the end users mentioned in the article as I think it’s a little unfair to give a marketing officer 3.5 hours to respond to something they were unlikely to be prepared for.
Hikvision rarely contribute as they engage in a ‘do not feed the troll’ approach and as is well documented, John and Hikvision rarely see eye to eye even when it’s plain to see they are both on the same side with the same intent.
Please will you rewrite the headline/article to be more factually correct? If you do not, it adds more flame to pre-existing belief that they are simply written to sell subscriptions!
If there is an issue with the product then that is an issue that lies at the feet of the manufacturer. If there is an issue with an installation then it is an issue with the installation which is either a training issue, citing here a lack of understanding of the product constraints or an installer who does not care.
In most cases, I do not think that the end user is at fault.
Why do you continuing beating on Hikvision when there are many more news worthy subjects?Why have you not included a beat up on Avigilon’s demo video that shows someone walking between the blackbody and the camera? Or Mobotix at European airports?or Dahua at the UK Nightingale COVID hospitals - surely that’s far far worse?
Agree
Disagree: 2
Informative
Unhelpful: 2
Funny
Create New Topic
Login to read this IPVM report.
Why do I need to log in?
IPVM conducts reporting, tutorials and software funded by subscriber's payments enabling us to offer the most independent, accurate and in-depth information.
Comments (21)
Undisclosed Integrator #1
@Charles - I feel the term ‘faulty’ is misused here, perhaps the term ‘incorrectly deployed’ is more correct?
Clearly there is likely an issue with the way these systems are installed and used.
There also seems like a training and sys op issue on some of these too...
There could be a robust process in place but your enquiries are likely to be received and answered by the marketing/press officer rather than someone with more knowledge?
Create New Topic
Mark Schweitzer
I would agree with my colleague, using correct language is important to credibility and the national inquirer style headlines I've seen around the subject lately undermine the valuable and well thought out reporting and writing that is actually done in the story.
To the article's point, it amazes me the number of end users out there that have spent lots of money on these "solutions" but did not do the proper research to make the decision or ensure they are properly being used. This laziness on the part of the customer is only amplified by the predatory tactics that some of those in this market appear to be using.
Create New Topic
Roy Nilsen
I think it depends how they are used and sold. Reducing Covid-19 is clearly directly related to mitigation strategies implemented. No single strategy (i.e. hand washing) is a panacea in itself. It is the combination of strategies that reduces, not eliminates, risk.
An infectious disease specialist explained it to us this way - think of each mitigation strategy (hand washing, social distancing and masks) as layers of swiss cheese. Individually they all have holes in them. But as you stack them together, the holes get smaller but not completely eliminated. I see elevated skin temperature detection as another layer of swiss cheese.
Manufacturers, dealers and end users need to be cautious of how they are promoting "fever" cameras. It should be promoted as part of an overall mitigation strategy as another element of mitigation. If St. Monica's Trust, for example, is promoting it specifically "to keep residents safe" that is irresponsible. But if they are implementing it with all of the aforementioned strategies, then I don't see a problem with it.
I think this is an important conversation to have though especially to protect manufacturer's from themselves. If such reporting can force manufacturer's to change how they are promoting, and be more responsible with their marketing, maybe dealers will follow? It would be a shame if a manufacturer just saw this as an attack and didn't look beyond themselves or their bottom line to actually see the point and adjust how they promote their solutions.
Create New Topic
Undisclosed Integrator #1
One question, how long do you give companies/people to comment before you publish?
Create New Topic